Fairy Tales

Project Hole Mary

Andy Weir’s “Project Hail Mary” through the eyes of an offended nuclear engineer and thermal physicist

A note for the reader. “Hail Mary” is not just a prayer. In American football, a Hail Mary pass is a throw at random: the last seconds of the game, nothing left to lose, the quarterback hurls the ball the length of the field and prays that somebody, anybody, catches it. That is exactly what the mission in the book is called: a desperate bet with a one-in-a-million chance. I suspect the book itself got its name the same way: it was hurled at the reader at random — with a hundred holes and stretches — and everyone prayed it would be caught. Judging by the sales, it was.

Contents

No, I Haven’t Read It

Let me confess right away: I had no intention of reading this book. The Wikipedia article was enough. My eyes started bleeding somewhere around the paragraph where a microbe flies from the Sun to Venus at 0.92 of the speed of light, brakes against the atmosphere, breeds, and flies back. I’m an engineer. My whole life I’ve been calculating where the heat goes. When someone tells me about an organism that absorbs heat and never gives any of it back anywhere, my eye starts twitching like an accountant’s who’s been handed a balance sheet with no liabilities side.

Then I sat down and read it anyway. Not for fun, but with a pencil: specifically for this article. To catch holes. I figured I’d collect a dozen, write an angry post for my page and calm down.

I ended up with 66 holes and over a hundred stretches. A hole is when the book contradicts itself or the laws it itself acknowledges. A stretch is when something could happen, but only if every character switches their brain off at the same time. The full register is in the appendix at the end of the article, numbered. It also marks which of these were already found on the internet before me and which, apparently, weren’t. The text only gets the juiciest bits.

Make yourselves comfortable. It’ll be long, it’ll hurt, there will be numbers. The second law of thermodynamics will leave the chat somewhere around the third chapter. Newton’s second law will follow, clutching its heart. The law of conservation of angular momentum will leave silently, without saying goodbye. The law of conservation of common sense never joined the chat in the first place.

I’m Not Against Science Fiction

Before I get filed under the nerds who “don’t get the genre,” a word in my defense.

I love science fiction. Harry Harrison sends his Stainless Steel Rat across half the galaxy and doesn’t waste a single line on how the interstellar drive works. It flies — good enough. Asimov writes about robots for forty years and never once explains what a “positronic brain” is. It’s positronic, period, move along, reader, the ethics are more interesting here. Edmond Hamilton writes about kings, princesses and star empires, and he has exactly as much physics as it takes to get the ship to the next chapter. Lem in “Solaris” honestly says: we don’t understand this ocean and most likely never will — and builds one of the best books of the century on that very incomprehension.

I have no beef with any of them. Because they follow the rule of the genre: one or two magic assumptions are declared up front, and from then on the world lives by its own logic and doesn’t cheat. The assumption is a canvas. On it you paint interesting questions: what is a human, what is a mind, what is power, how will an engineer behave if tomorrow everything breaks.

What you shouldn’t do is bury yourself under smart words. If you write “positronic brain” and shut up, I’m fine. If you write “Astrophage has mitochondria, ATP, DNA, it stores energy in neutrinos, glows at 25.984 µm and holds a temperature of 96.415 °C,” you’ve invited me over with a calculator yourself. You yourself said: “Check me, I’m scientific.” Fine. Checking.

There’s one more rule that everyone who’s ever written a plot knows: no pianos in the bushes (no deus ex machina). If the hero is cornered, the way out must be planted in advance, and honestly. If in the climax a piano rolls out of the bushes, the author has lost. This book has so many pianos that you could rent the bushes out to the philharmonic.

The Sequel Nobody Ordered

Look at how Hollywood sequels work. In the first “Alien” there’s one alien. In the second, a queen and hundreds of soldiers. In the first, a space truck with tired truckers. In the second, a planetary base. Instead of a crew, a SWAT team. Instead of a cat, a whole little girl. More of everything, louder, more expensive.

“Project Hail Mary” is literally “The Martian 2.” Before: one man stuck on one planet. Now: one man stuck at another star. Before: survive. Now: save the Sun. Before: potatoes in his own shit. Now: a whole alien microflora. Before: NASA. Now: the whole Earth under the command of one woman with unlimited powers. Before: the hero cracks jokes for the camera. Now: the hero cracks jokes for an alien. Before: “I’m gonna have to science the shit out of this.” Now: same thing, except the science no longer fits the format.

And here’s the trap. “The Martian” was good precisely because the physics there was almost real: pressure, oxygen, calories, hydrogen. Weir did the math, and the reader was in heaven. Competence porn in its purest form. The second time around the author decided to repeat the high, but the scale grew by twenty orders of magnitude while the calculator stayed the same. It came out like a magician who pulls a rabbit out of a hat at a kids’ party and then tries the same move in Las Vegas with an elephant. The elephant doesn’t fit. The hat is cracking. The audience, for some reason, applauds.

Why They Applaud

Because they applaud. On Goodreads the book has 4.51 out of 5 from almost two million ratings, 90% give it four or five stars. Bill Gates and Barack Obama recommend it. The 2026 movie with Ryan Gosling has 95% on Rotten Tomatoes. NASA provided logos, an astronaut consultant on set, a screening on the ISS and a panel at JPL.

But look more closely at who exactly is praising the science. NASA is a promotional partner, that’s normal PR, there’s no scientific review in a press release. Astronauts praise it for being “inspiring.” Writers on the cover praise “lots of science in the science fiction.” But the actual scientists whom journalists did bother to ask are cautious: “internally consistent,” but Astrophage’s energy is off “by orders of magnitude,” you can’t survive on the Sun, a multi-year coma means brain damage. Neil deGrasse Tyson, on air with Weir himself, corrected him about a planet’s magnetic field. Writer Mary Robinette Kowal in the Washington Post called it “refrigerator logic” — the inconsistencies you notice once you’ve already closed the book and gone to the fridge.

I see three explanations for the raptures, and I like none of them. One: hype — the marketing budget of the movie and the publisher works like a pump. Two: the reader’s level has dropped — for most people “neutrino” and “mitochondria” sound like “positronic brain,” and it never occurs to anyone to check them. Three: the genre is in crisis, and with slim pickings anything with formulas in it gets called “hard” SF. Most likely all three are true.

Now about how it’s written. Flat characters: the protagonist is a Mark Watney clone, the supporting cast are national stereotypes (the Russian with vodka, of course). Eighth-grade humor. A physics textbook dissolved in dialogue, and a textbook with mistakes in places at that. The hero’s amnesia as a storytelling device — a cliché people were already writing about back in 2021. And the main thing: every time the plot needs it, the world changes its rules. That’s what the rest of this is about.

Rules of the Game: Astrophage Exists

I’m a fair opponent. I accept the book’s main assumption in full, no haggling.

There’s a microbe, Astrophage. It absorbs light and heat, turns the energy into neutrino mass and stores it inside itself. When needed, it releases this energy as a directed beam of infrared light and flies on it like a photon rocket. Energy density on par with antimatter. Its cell holds a temperature of 96.415 °C, not a degree more. It lives on the surface of the Sun, breeds in the atmosphere of Venus where there’s CO₂, and shuttles between them. There’s gotten to be so much of it that the Sun has dimmed. That’s it. Accepted. Neutrinos, membrane, “super-cross-section” — fine. This is my Harrison, my positronic brain.

From here on I don’t touch the physics of Astrophage itself. I touch the logic of the world it exists in. And this is where the fun begins: the book, having accepted its own assumption, immediately breaks it. It’s not me breaking its world — it breaks it itself, every time the plot needs it to.

Chapter 1. The Creature That Eats Heat, or The Second Law Has Left the Chat

Why Fly to the Sun If You Don’t Have To

Let’s start with the most important question, which for some reason not a single character in the book asked themselves, including the scientists of two civilizations.

Weir himself, in three different places, explains to us that Astrophage doesn’t need a star to charge up, it needs any heat above 96.415 °C. The Eridians breed it in metal spheres dropped into their 210 °C ocean. Earthlings breed it in the Sahara under glass. The hero spells out the instruction directly: give it “lots of heat energy.” That’s it. The Sun is optional. A heat source above 96 °C will do.

Now open any reference book on the atmosphere of Venus. Below roughly 47 kilometers the temperature there is 100 to 143 °C. Carbon dioxide: 96.5%. Infrared emission of CO₂ at those very 4.26 and 18.31 µm lines that Astrophage supposedly responds to — as much as you want, it’s coming from every direction. In other words, in the clouds of Venus Astrophage has everything at once: heat, carbon, a breeding trigger. Just sit there and divide.

What does a settled life cost? For a cell to charge up on the spot it needs on the order of 4·10⁻⁷ joules. A round trip to the Sun at 0.92 of the speed of light costs it about 1.5·10⁶ joules per cell. The difference is a factor of four trillion. An evolution that, over billions of years, produced an organism spending four trillion times more energy than necessary on a trip it doesn’t need isn’t evolution. It’s a government contract.

The optimal Astrophage in Weir’s own world would sit in the clouds of Venus, quietly breed there and leave the Sun alone. No dimming. No mission. No book. The plot rests on the microbe being dumber than the author.

The Thermal Engineer Takes the Floor

Now I’ll take off my glasses and speak as someone who has spent his whole life calculating heat pumps, refrigeration cycles and reactor loops.

You can’t just “eat” heat. Heat is the lowest-grade energy in the Universe. To turn it into something useful — work, electricity, neutrino mass with the energy density of antimatter — you need a temperature difference. Heat has to flow from hot to cold, and only part of that flow can be taken. That’s the second law of thermodynamics, that’s the Carnot cycle, that’s what every power plant, every refrigerator and every combined heat and power station stands on. Without a cold end, nothing works.

Astrophage holds 96 °C. The Eridians drop it into an ocean at 210 °C. Heat flows into the cell — fine. But where does the cell dump its “exhaust”? There’s nothing colder than 96 °C around. The sphere is closed. The ocean is hot. The cell takes heat from a hot body and stores it as perfectly ordered energy, giving nothing back anywhere. That’s a perpetual motion machine of the second kind. Patent offices have been rejecting such applications without review for about a hundred and fifty years.

Same with the Sahara, only funnier. According to the book, the panels yield about 0.5 micrograms of Astrophage per square meter per day. Convert via E = mc²: that’s about 520 watts from every square meter, around the clock. The Sun in the Sahara gives on average roughly 250–300 W/m² over a day, and that’s light, not electricity. Even if you assume the cell works as an ideal heat engine between a hot panel and a cold sky, the real output comes out 8–15 times lower than claimed. The book didn’t just round off. The book is selling efficiency above one hundred percent.

The second law of thermodynamics has left the chat. Didn’t slam the door — just quietly left, the way an engineer leaves a meeting room when the marketing guy starts drawing arrows on the whiteboard.

The Brightest Venus in the Universe

How does Astrophage find Venus? According to the book, by the infrared light of carbon dioxide. It sees the CO₂ lines and flies there. Thrust proportional to the strength of the IR signal.

Okay. Now a question for an A: who in the Solar System shines brightest at 4.26 µm? Correct, the Sun. At this wavelength the Sun is brighter than Venus by roughly a factor of 4·10¹¹. If a cell flies toward the brightest CO₂ IR source, it flies into the Sun. Always. No options. It simply won’t notice Venus next to a searchlight.

Weir apparently sensed this, and so he adds a magnetic stage: first the cell leaves the Sun’s pole along the field lines, because the Sun is “too blinding,” and only then looks for Venus. Fine. Except in one chapter this magnetic start is mandatory, while in another — in Grace’s lab, in the ship’s spin drive, in the scene at Adrian — the cells switch on perfectly well from IR alone, no magnets needed. The rule changes depending on what the scene needs.

And on top of that, the Sun’s field lines in a polar coronal hole diverge super-radially, fanning out. At a distance of 37 million kilometers, cells following them would form a cone tens of millions of kilometers wide, not the thin red thread Petrova saw in the telescope. And where, may I ask, is the same line from the south pole? And what happens every 11 years when the Sun flips its polarity?

How to Aim at Venus When You’re Smaller Than the Wavelength

An Astrophage cell is about 10 microns. The signal wavelength is 4.26 and 18.31 µm. Anyone who’s ever held an antenna knows: a receiver a couple of wavelengths across can’t determine direction more precisely than to within tens of degrees. The diffraction resolution of such a cell is about 24° at 4.26 µm and about 105° at 18.31 µm. In other words, at the long wavelength the cell can’t even tell whether the signal is in front or to the side.

From the top of the arc, Venus subtends about 22 arcseconds. That’s four thousand times finer than the cell is able to see. Over the whole flight, on the order of twenty photons of the right frequency reach the cell from Venus. Twenty. Try homing in on a lighthouse when you’ve seen twenty flashes the entire way and your sight is the size of half the sky.

And that’s without Doppler. At 0.92 c, a cell flying toward Venus sees the CO₂ lines not at 4.26 and 18.31 µm but at roughly 0.87 and 3.74 µm. Its “eye” is tuned to one frequency, and the target is shining at a completely different one. The cell needs a tunable detector. With DNA and mitochondria, sure.

Braking: Ten Billion g

According to the book, the Indian agency ISRO measured the speed of Astrophage between the Sun and Venus: 0.92 of the speed of light. Both ways. That’s not my estimate, that’s canon.

At that speed the cell arrives at Venus and “brakes in the atmosphere.” The atmosphere of Venus is about 300 kilometers thick. At 0.92 c the cell crosses it in about two milliseconds. Killing 0.92 of the speed of light in two milliseconds is a deceleration on the order of 10¹⁰ g. Ten billion Earth gravities. Pressure on the cell: from single digits to tens of gigapascals. The cell won’t get crushed. The cell will get turned into plasma, and the plasma will keep flying.

The power of all this braking at Venus is about 4·10²¹ watts. That’s two hundred million times the entire energy output of humanity, and all of it released in the upper atmosphere of one planet. Venus should be blazing in gamma rays so hard that Fermi-LAT would see it as one of the brightest sources in the sky. In the book — silence.

And the way back? To leave Venus back toward the Sun at 0.92 c again, the cell needs a mass ratio that eats up almost everything it’s accumulated. To turn 109° at the top of the arc — as drawn by Petrova — the cell at that speed would effectively have to stop and accelerate all over again, that’s a mass ratio of about 576. The “rose slowly and then bolted” scheme doesn’t explain why the entire arc is visible.

And then, according to the book, after division both daughter cells fly to the Sun. So the return flow should be twice the outbound one. Petrova should have seen not one arc but a triangle: the arc out, the chord back, and the return leg brighter. One arc is visible. The law of conservation of mass has followed the second law of thermodynamics out the door.

The Nuclear Engineer Is Furious: The Word “Neutron”

Now I’ll speak as a nuclear guy. I specifically searched the English text of the book for the word “neutron.” Zero matches. Not a single one. In a book about particles flying at 0.92 c, about radiation, about nuclear bombs and about energy on the scale of E = mc².

What happens when a proton, a carbon or an oxygen nucleus at 0.92 c slams into an atmosphere made of CO₂? Nuclear spallation reactions. The nucleus shatters into pieces, out fly neutrons, protons, light nuclei. With the Astrophage flux needed to dim the Sun by 10%, braking at Venus would produce on the order of 10³² neutrons per second. At ISS-orbit altitude above Venus, the background would be on the order of ten millisieverts per hour. There’d be hundreds of thousands of times more carbon-14 in Venus’s CO₂ than natural levels. Cells that passed through this hell would be radioactive: a milligram of Astrophage would carry on the order of a hundred becquerels from carbon alone, and from tritium, which forms during spallation of oxygen and carbon, 10⁸ becquerels and up.

Tritium inside a watery cell means about 0.1 gray per second. Ten thousand grays a day. The most radiation-resistant bacterium on Earth, Deinococcus radiodurans, withstands on the order of sixty grays per hour. Secondary particle cascades at about 50 km altitude give thousands of grays per second. The breeding layer at Venus is sterile. Nobody can breed there. That’s it, the book ended at the discovery chapter.

Oh, right. According to the book, killing Astrophage takes a nanoneedle. But a relativistic proton passing through the cell at 0.92 c and leaving millions of grays per second in its wake somehow doesn’t scare it. A needle — yes. A nuclear shell — no. It’s like a person who dies from a pinprick but calmly survives being shot by a tank.

Smart Words That Bury the Author

Remember, I said: don’t say smart words and there’ll be no criticism. Here’s an example where the author dug his own grave.

In the book, Astrophage mitochondria are 98% identical to Earth ones. The hero and his colleagues conclude: common ancestor, panspermia, life came from space four billion years ago. The problem is that the mitochondrial DNA of humans and chimpanzees, which split only six or seven million years ago, differs by about 9%. Four billion years of separate evolution with 98% identity is impossible. 98% means recent contact. Very recent. Remember this, we’ll need it in the last chapter.

Or: “there are no elements on the Sun other than hydrogen.” Not true, there’s helium, carbon, nitrogen, oxygen, iron — everything in the periodic table. But here’s the funny part. If there were only hydrogen there, where does Astrophage get nitrogen and phosphorus for its DNA and its ATP? Which the author wrote into it himself. The word “ATP” means phosphorus. The word “DNA” means phosphorus and nitrogen. You said “A” — now pay for “B.”

Speaking of phosphorus. If you calculate how much phosphorus an Astrophage population dimming the Sun by 10% needs, you get 3·10¹¹–3·10¹² kilograms per year. The clouds of Venus hold, by the most generous estimates, enough phosphorus for weeks or decades of such consumption. The population ceiling isn’t “10% of luminosity,” it’s Venus’s phosphorus reserve, which the cells haul off to the Sun and never bring back. Incidentally, this would explain the mysterious 10% plateau everyone in the book racks their brains over. But nobody in the book thought of it.

The Lab Episode: The “Dark” Cabinet

I particularly love the lab episode. Grace figures out that Astrophage responds to the CO₂ lines (4.26 and 18.31 µm) and that in a dark cabinet the cells go still.

A dark cabinet. At room temperature. Any body at 20 °C glows in the infrared — and precisely at the wavelengths in question; the cabinet walls shine at 4.26 and 18.31 µm brighter than Venus from the distance at which the cells are looking for it. If the cells responded to these lines, they’d be bouncing from wall to wall in the cabinet. They go still. So the cells are responding to something else, and the book’s whole “scientific” chain is fitting the data to the answer.

And the cherry on top: the glass bulb of the lamp Grace shines on the cells to play Venus is opaque to radiation longer than about 5 µm. So 18.31 µm doesn’t get through it at all. The “experiment” had only one of the two stimuli, and the conclusion was drawn about both.

Chapter 2. The Sahara Under Glass, or How Humanity Picked the Worst Option of All

A Quarter of the Sahara

So, the Sun is dimming, and in twenty years half of humanity will starve to death. Earthlings need fuel for a starship — two million kilograms of Astrophage. What does united humanity do under the leadership of a woman with unlimited powers?

It covers a quarter of the Sahara with solar panels. Two trillion square meters. Each panel is black anodized foil, glass with a one-centimeter gap, a magnet, an IR filter and a little hole for CO₂. Astrophage grows inside. A former convict, sentenced for the deaths of seven people at a solar tower and for embezzlement, promises a ton a day, and the fuel will be ready “in a few years.”

Let’s count the glass. Two trillion square meters takes 6 to 23 gigatons of glass, depending on thickness. All of humanity produces about 0.1 gigatons of glass a year. So you need 60 to 230 years of global glass production. In “a few years.” Without logistics, without railroads across the desert, without water desalination, without housing.

Next, operations. The Sahara is dust. Panels need cleaning. Dry-cleaning robots means one to two billion robots. Opening the panels once a year to harvest means five and a half billion panels a day and eleven million people on harvesting alone. Plus replacements. Plus repairs. Staff: ten to thirty-five million people with full robotization, hundreds of millions by hand. The book devotes exactly one sentence to all of this: the farm “regularly delivered as promised.” Thanks, very informative. In my life I’ve commissioned facilities where a single pumping station came with three volumes of operating documentation.

And one more detail that fills me with quiet joy. The Sahara farm harvests solar energy from Earth and stores it in Astrophage, which then flies away on a starship. The Sun is already dimming, Earth is already short on heat. And we take another 10¹⁵ watts or so off the surface and ship them to Tau Ceti. That’s about 9% of the deficit the dimming created. In other words, the rescue operation itself adds a tenth of the disaster to the disaster. It’s like fighting a fire by pumping water out of the fire reservoir into the next county.

Now Look at Venus

Somewhere in those same years, in this same book, the ArcLight probe has already been to Venus and collected Astrophage samples there. So people already know how to catch Astrophage at Venus. And they know there’s a stream of it there. How much?

If you estimate the flux needed to dim the Sun by 10%, on the order of two and a half million tons of Astrophage pass Venus every day. Two million kilograms for the Hail Mary is one or two minutes of that flow. Not years. Not a quarter of the Sahara. Minutes. A tanker that flies to Venus, brakes in the atmosphere, scoops from the breeding layer at 50–60 km and comes back to Earth would spend about 0.4% of its mass on the whole trip — because it has, excuse me, an E = mc² engine.

This isn’t my genius idea. It’s an idea lying right on the surface of the book. Any first-year student could have suggested it. Nobody suggests it. The scientists of the entire planet, led by the most effective manager in human history, stare at a free gas station orbiting the neighboring planet and build an oil refinery in the desert.

Let’s go further. If Astrophage feeds on sunlight, we’ve got the Sun right next door — and that, mind you, is a living Dyson sphere. A population intercepting even 0.1% of the Sun’s luminosity is about 4·10²³ watts. Twenty billion times all of humanity’s energy. A free battery with the energy density of antimatter that charges itself and delivers itself. In the book they diligently wipe it out completely, down to the last cell. One character once mutters that it’d be nice to farm Astrophage on the Moon — and nobody ever brings that line up again.

And the cherry on top. Earth has an engine with an energy density of E = mc². To put a ton of cargo into orbit on a thermal rocket running on Astrophage, you need on the order of 0.4 milligrams of fuel. Milligrams. Yet the Hail Mary is assembled in orbit from modules launched by sixteen ordinary rockets running on kerosene and hydrogen. It’s like inventing teleportation and still delivering the mail by horse because “that’s what the regulations say.”

Where’s the Target?

Okay, fine, we have fuel. We fly to Tau Ceti. Why there? Because Tau Ceti is a star in the “infected” zone that for some reason isn’t dimming. So there’s something there that kills Astrophage. Logical?

Almost. Tau Ceti sits in the middle of an infected cloud of neighboring stars. It isn’t dimming. There are at least three explanations: (a) there’s a predator there; (b) Astrophage didn’t take hold there; (c) Tau Ceti is the source, the oldest outbreak, and it’s been sitting on the minus-10% plateau for ages, and we simply don’t know what it used to be like. Option (c) isn’t discussed on Earth at all. Yet it’s the first one any epidemiologist would think of: if everyone around one point is sick and that point is “healthy,” that’s patient zero, not miracle immunity.

Humanity pours all its resources into one mission to one star without checking whether the target makes sense. It’s like betting the planet’s life on roulette without checking whether your number is even on the wheel.

There’s no Plan B. There’s no parallel program to find an Earth predator for Astrophage — even though the book itself shows that Astrophage can be farmed in unlimited quantities, and later Taumoeba gets trained to tolerate nitrogen in 23 generations. The Astrophage genome is read at the very start of the book — “Analyzing its DNA. Good for them,” — and that’s it, nobody ever mentions it again. No phages, no antibiotics, no genetically modified farm strains, no genetic weaknesses. Genetics has been invented but isn’t used.

Meanwhile, in Antarctica they set off 241 thermonuclear charges to release methane and warm the planet. A noticeable effect would need hundreds of gigatons of methane. Charges 50 meters below the surface aren’t “minimal radiation,” that’s a textbook cratering explosion with radioactive fallout over half the Southern Hemisphere. The word “neutron,” let me remind you, doesn’t appear in the book.

Power Without a Counterweight

Why is everything like this? Because Eva Stratt is in charge. She was granted immunity “for any crime anywhere on Earth.” She decides everything: crew composition, mission profile, accident investigations, who lives, who dies. She has no council, no audit, no opponents.

And that’s exactly why every decision in the book gets made as if by one person with tunnel vision. Fuel — in the Sahara, because that’s what was decided. Ship — one, because that’s what was decided. Target — Tau Ceti, because that’s what was decided. No communication with Earth, because nobody thought of it. No Plan B, because there’s no time. The discoverer of the line, Irina Petrova, is nowhere to be found in the project or the working group — even though the line, the problem and the instrument are all named after her. They just forgot about her.

This, by the way, is the one place where the book could have posed a genuinely interesting question: what happens to humanity when it hands all power to one person for the sake of salvation? The Manhattan Project, incidentally, was exactly about that, and it had Oppenheimer, and Groves, and Szilard with his letters against the bomb. The book doesn’t pose the question. Quite the opposite: “Her plan worked perfectly.” Coercion approved. Price not named. Moral: if a sociopath is running everything, everything will be fine.

Chapter 3. The Powder Keg, or The Nuclear Engineer Clutches His Heart

A Jar Holding a Photon Bomb

Here I’ll tell you something everyone who’s worked near fissile materials knows. A nuclear weapon has several independent interlocks. Electrical, mechanical, logical, spread across different subsystems. For it to go off, all of them have to trigger at once, in the right order and on a signal that can’t be forged. That’s not paranoia. That’s decades of experience of bombs falling off planes, burning in fires and getting into accidents — and not exploding, because they were designed not to explode.

Now the Hail Mary. In the tanks: two million kilograms of Astrophage paste at 96 °C. CO₂ is dissolved in the paste. Astrophage fires its engine when it sees CO₂ infrared light. And any body at 96 °C glows in the infrared. The paste is its own trigger. It’s always armed. One fed cell that fires releases about 1.5 megajoules — enough to heat four kilograms of neighboring paste by 90 degrees. The neighboring paste starts glowing brighter. And so on.

Safety catches? None. No compartmentalization into cells, no cooling below threshold, no magnetic arming, no genetic key strain. Nothing. The tanks hold energy on the order of 2·10²³ joules. That’s fifty thousand gigatons of TNT equivalent. For comparison, humanity’s entire nuclear arsenal at the peak of the Cold War was about twenty gigatons. The Chicxulub asteroid that killed the dinosaurs was on the order of a hundred thousand gigatons. We’re in the same weight class.

And this thing is hanging in low parking orbit. In a decaying orbit with a 51.6° inclination, over every inhabited latitude, with a margin of a few weeks. It could climb higher in minutes — the ship, let me remind you, has a spin drive. Nobody raises it.

And then it launches. From low orbit. Straight out. The spin drive exhaust during acceleration is about 8·10¹⁹ joules, about twenty gigatons of TNT equivalent, and at the start of the burn part of that beam points straight at Earth. Tangential thrust, spiraling out, a preliminary climb to a high orbit — none of this is even considered in the book. Humanity, saving itself from freezing, risks burning off its own atmosphere.

Baikonur: A Milligram Instead of a Nanogram

Now an episode that should go into safety textbooks as an example of how not to do it.

A few days before launch at Baikonur, scientists are testing a rare generator failure. The procedure calls for less than a nanogram of Astrophage. By mistake they’re issued a milligram. A million times more. A milligram of Astrophage is 9·10¹⁰ joules, about 21 tons of TNT. Explosion. DuBois, Shapiro and fourteen other people die.

Question number one: a milligram and a nanogram are a visible speck and an invisible nothing. You can see the difference with your naked eye. How do you mix them up? Only if the substance is stored without labels, without weight control, without a second signature. In a lab working with antimatter-grade energy.

Question number two: who was in the room? The mission’s primary science specialist and his backup. Together. At a destructive test with fuel. A few days before launch. That same Stratt who, according to fans, wouldn’t let the prime and backup crews fly on the same plane. Throughout the entire aviation and space industry, prime and backup crews are kept apart. Tests are run by test engineers, not by the people you’re about to send to a star. Here — a pair in one room, setting up a dangerous experiment themselves. A common-mode failure is built into the architecture.

Question number three: who investigates the accident? Stratt. Personally. In two days. Without Grace. A person with motive, opportunity and full control over all the data delivers her own verdict: “dosage error.” I don’t want to insinuate anything. Not yet. The chapter about the hypothesis comes at the end.

“A Muffled Whump”

A spin drive test in a hangar. Thrust of 60,000 newtons for 100 microseconds. 1.8 gigajoules of energy go into the target — a ton of metallic silicon. About 20 micrograms of Astrophage burn. The numbers, by the way, are consistent: F = P/c. Here Weir did the math right.

And then he writes: “muffled whump.” And a ton of silicon melted.

Guys. 1.8 gigajoules in 100 microseconds is a power of 18 terawatts. A beam that powerful won’t have time to heat a ton. It’ll be absorbed in a thin surface layer: twenty to two hundred grams of silicon will instantly turn into plasma. That’s an explosion of about 0.43 tons of TNT. In a closed hangar. That’s not a “whump,” that’s the hangar doors blown out and everyone inside concussed. And melting a ton of silicon takes about 2.9 gigajoules — more than there was energy in the entire pulse. It wouldn’t have melted.

One more detail. This hangar is on an aircraft carrier. The engine under test held two grams of Astrophage, of which twenty micrograms burned. Two grams is 1.8·10¹⁴ joules, about 43 kilotons. Three Hiroshimas. On a ship with a crew of five thousand. In the form of paste at room temperature, which, as we remember from the dark cabinet, glows at itself in IR.

How the Engine Works

According to the book, the spin drive works like this: Astrophage paste is spread on transparent panels that rotate like a revolver cylinder, and CO₂ light is fed through the panel to switch the cells on. First, the panel: in the book it’s polycarbonate. Polycarbonate is opaque in the mid-IR, exactly at 4.26 and 18.31 µm. The trigger won’t get through. Okay, fine, make it sapphire. But any real material absorbs at least something. Parasitic absorption of one millionth of a single element’s 9 terawatts is 9 megawatts in the panel. It’ll vaporize.

Next: a fed cell reflects light (that’s in the book, p. 110). So cells in a thick layer of paste will reflect their neighbors’ light back into the engine. You need a monolayer. Even a perfect monolayer sends back a noticeable fraction of the power — about a terawatt per element, with a tolerance that in the book must be under one billionth. The oil the paste is thinned with evaporates right into the beam and scatters it. Soot and burnt-out cell membranes are “scraped off by a scraper.” With a scraper. Out of a beam of terawatts.

Alone, in ten minutes, in my head, while walking the dog, I sketched a working version: a tape cassette with a monolayer of underfed cells on a gold mesh, raster illumination in a “checkerboard” pattern so that fed cells are surrounded by half-starved ones that absorb their neighbors’ parasitic heat, and side illumination for the trigger. This solution closes almost all the problems. Ten minutes. The dog, by the way, didn’t object either. The characters in the book had several years, an unlimited budget and the entire scientific potential of Earth. What they have: one design, one test lasting 100 microseconds and not a single endurance test. And off we fly.

The Photon Rocket Equation

One last thing about the ship. By the book’s numbers, the Hail Mary flies at 1.5 g for about four years of proper time — accelerating and decelerating. For a photon rocket, the ratio of initial to final mass for that is about 415. Of the 2.1 million kilograms at launch, no more than five tons reach the target. That’s less than the habitat module with the centrifuge weighs on its own, never mind the lab, the tanks, the Beetles and Grace himself. The ship can’t brake at Tau Ceti. The ship described in the first chapter of the book doesn’t make it. That’s not my “hypothesis.” That’s the Tsiolkovsky equation in relativistic form, which the author, by all appearances, checked exactly once — for the acceleration leg.

Chapter 4. The Crew, or Suicide Squad

One Ship, Three People

A mission to save humanity. The cost of failure: the death of half the people on Earth. How many ships? One. How many people? Three. All different specialties, no redundancy. If one dies, nobody can do his job.

Next: all three fly in an induced coma, because a multi-year flight while conscious would break them. Only carriers of rare genes survive the coma, about one in seven thousand. Studies showed that carriers survive the coma in one hundred percent of cases. In flight, one of the three survived. Why? Nobody in the book asks that question. If my rocket delivered 33% successful launches against a demonstrated test reliability of 100%, I’d start the review not with “who’s to blame” but with “what didn’t we measure.”

Two ships. At least two. That’s standard practice in every space program where the cost of failure is high: the Vikings flew as a pair, the Voyagers as a pair, even the movie “Armageddon” had two shuttles. The irony is that they built four Beetles — the capsules meant to carry the answer back to Earth. There’s redundancy for the flash drive. For the people — none.

The ship was assembled from modules over sixteen launches, the bottleneck was fuel. But we already know: fuel could have been taken from Venus in minutes. So the second ship isn’t limited by fuel, it’s limited by willingness. A second set of modules could have stood on Earth as a ground test rig and then gone out after it. Ahead of the ship you could have sent a Beetle-class scout — for a fraction of a percent of the fuel — to Tau Ceti and along the route, to find out whether there’s anything to look for there at all, and whether there’s a dust cloud on the way. A one-microgram dust grain at 0.9 c carries a hundred megajoules. None of this exists.

A Coward Under Anesthesia

And here’s the culmination of the planning. Three days before launch, the primary science specialist and his backup die. In their place Stratt appoints Ryland Grace — a schoolteacher, former molecular biologist, who wrote a paper arguing that life doesn’t necessarily need water. By this point his thesis has already been refuted: Astrophage is mostly water. He hasn’t done science in years. He refuses to go. Stratt tells him: “You’re a coward.” And then they knock him out and send him by force, with his memory wiped.

Humanity sends to a star a confrontational man who didn’t want to go. In a crew of three. In the only ship. With no backup briefing in case of amnesia, even though Stratt knows about his amnesia because she ordered it herself. She lies to Yáo, the mission commander, that Grace asked for the sedation himself. The commander, by the way, was against coercion.

Try to imagine you’re the commander. You wake up at Tau Ceti and find next to you a man who didn’t want to be here and was brought by force. You don’t have a gun. Yáo, by the way, asked for a gun. This is a recipe for sabotage, suicide or murder. The book declares all of it a good deal.

Oh yeah, and the age. According to the book’s chronology, Grace is about 29 at launch. So at recruitment he was 23–26. In that time he managed to defend a PhD, write a scandalous paper, get kicked out of science and work as a teacher for several years. The discrepancy is five to ten years. Apparently that got wiped too.

Jack of All Trades

Next, Grace on his own — a biologist by training, a teacher by profession — does relativistic trajectory calculations, goes on spacewalks, writes software to recognize the chords of an alien language, does chemistry, nuclear physics, navigation and medicine. Astrophage, meanwhile, remains a black box to him: no genomics, no biochemistry, no physiology. He’s a biologist who spends the whole book doing physics. The microscope work and the Taumoeba selection are the level of a good undergraduate lab course.

Grace’s qualification is a self-fulfilling prophecy. He was picked for a thesis that had already been refuted. He got mission training because Stratt kept him attached to the project for years in order to send him at the right moment. Meanwhile there was a volunteer chemist whom nobody trained. The qualification wasn’t found, it was assigned.

Amnesia by Screenplay

Grace’s amnesia comes from a drug administered on Stratt’s orders via the ship’s medical bed. But this amnesia has a strange profile. He’s forgotten his own name, but remembers physics, chemistry, the English language and how to use the equipment. That’s not the profile of drug-induced amnesia but of dissociative amnesia — when a person suppresses trauma. And the memories come back to him as ready-made scenes, in the order and at the moment the plot needs them. Not earlier, not later. Very convenient amnesia. It doesn’t exist, but it’s very convenient.

And another thing: the crew wakes up and there are no instructions. No “read me if you don’t remember anything” note. No video. No file on the computer. The ship’s computer, by the way, also fails to warn about an approaching alien ship. The mission was made stupid so the protagonist would look smart — I didn’t come up with that, Mary Robinette Kowal did, in the Washington Post. Inside a starship, a person with amnesia who doesn’t know the procedures is a dead man walking. The first mistake could be the last. In Lem’s “The Invincible,” people with wiped memories are the first thing that kills the expedition.

Burial in the Exhaust

Grace sends his dead crewmates into the airlock and jettisons them. Under thrust. Without tissue samples, even though whether he should go into a coma for the return flight depends on what killed them. The cause of death, by the way, is never established. Nobody investigates anything.

Now about the airlock. The ship is flying under 1.5 g of thrust. A body jettisoned from the ship falls behind — that is, it flies backward, to where the terawatt-class spin drive exhaust is. The beam vaporizes the body in a millisecond, the plasma becomes a mirror and reflects 10¹³–10¹⁴ watts back onto the stern. With a tolerance that for the engine has to be under one millionth of its power. That’s enough to burn out the engines. There’s no interlock on the “Jettison” button while the engine is running. A man with no memory, who doesn’t know you’re not supposed to litter, throws bodies overboard. The ship survives it. Why? Because the chapter ended.

Chapter 5. Adrian, or Newton Has Left the Chat

If you thought it couldn’t get any worse — welcome to planet Adrian. This is the episode fans call the most engineering-heavy in the book. Grace and Rocky descend into the planet’s atmosphere to fish a sample on a ten-kilometer chain out of the layer where Astrophage breeds and, they hope, where its predator lives. This is the best episode in the whole book for number-crunching, because here three of Newton’s laws, the law of conservation of angular momentum and the law of conservation of common sense all break at once.

Hovering on a Photon Engine

The ship hovers in Adrian’s atmosphere, tilted at 30°, on the thrust of a photon engine of about 5.7·10¹⁴ watts. The beam slams into the atmosphere. The atmosphere reflects and scatters part of the light back, the white-hot air glows. According to the book, this reflected light melts a twenty-by-ten-meter hole in the ship’s port side over an hour of hovering.

Okay, the hull melted in an hour — that’s about right, order-of-magnitude-wise. But there’s a question here that for some reason nobody asked. What else does this reflected light see? The engines themselves. Specular reflection off the plasma and gas sends back 10¹¹ to 3·10¹² watts, the glow of the heated air another 4·10⁹ watts or so. An engine that works on the “cell fires on IR” principle gets hundreds to thousands of times more IR than it can tolerate. The spin drives should have failed within seconds or minutes — either by flashing off all the fuel or by overheating the panels. Instead, the ship got one side melted and flew on. The 30° tilt, by the way, doesn’t divert the reflection: scattering in the atmosphere goes in every direction.

The Chain That Knows Which Way to Hang

Now the chain. Ten kilometers, two hundred thousand links, with a sampler at the end. The ship hangs at 30° to the vertical, moving at about a hundred meters per second, and the chain hangs below almost vertically, deflected by 2°.

Guys. The ship hangs because engine thrust balances gravity. Thrust is directed along the ship’s axis. If the ship is tilted at 30°, the thrust has a horizontal component, and the ship accelerates sideways — and air drag pulls the chain back, toward the exhaust. A chain suspended from a ship held up by thrust hangs toward the exhaust, not vertically. At 30° and a hundred meters per second the chain is almost horizontal: deflection from vertical is 82–88°. And if the chain hangs at 2°, as in the book, it hangs right in the beam. Then the very first links vaporize, and once again we get a plasma mirror reflecting terawatts onto the stern. Thirty degrees, a hundred meters per second and two degrees do not coexist. Pick one.

And we still haven’t asked the main question: why a chain? Ten kilometers of links is the worst possible solution of them all. A chain is heavy, it rattles, it swings, every link is a separate hinge, every hinge is a separate point of failure, and you can only drop it one link at a time. The drillers from “Armageddon” are nervously chain-smoking in a corner: they can’t understand how anyone could fail to think of screwing rods together. A rigid string of pipes threaded one into the next is a technology Earth has had for more than a hundred years. You can hold it at an angle, the exhaust doesn’t blow it away, you can reel it back in.

Better still, a cable. And here it turns out that “naive Rocky” has completely forgotten something: his civilization has long known how to make xenonite cables. The Eridians have a space elevator made of xenonite, from the equator to synchronous orbit. Thousands of kilometers of cable holding its own weight in a two-g gravity field. And for a ten-kilometer sampler, Rocky forges a chain. According to the book, Rocky is an engineer. According to the book’s logic, Rocky is a talking CNC machine: Grace writes the spec, Rocky executes it, never proposes anything himself and never remembers his own planet’s elevator. For an engineer, that’s an insult. I am insulted.

Links Flying Into the Beam

Next, the chain had to be dropped. According to the book, it’s dropped link by link. Two hundred thousand links. Where do links go when detached from a ship hanging on thrust? Correct, they fall against the thrust. That is, into the beam. Each link, hitting the beam, vaporizes in a tenth of a millisecond and sends about 10¹² watts back onto the stern — roughly like twenty-five kilograms of TNT going off right at the nozzles. Two hundred thousand times. If I were this ship’s spin drive, I’d quit.

The Centrifuge: Grace’s Law

After that the ship has to be spun up to restore artificial gravity on board. According to the book, when the centrifuge cables are let out, the acceleration in the habitat module drops by a factor of 14, and Grace derives the law by which it all works.

The law is wrong. When cables are let out in free rotation, what’s conserved is angular momentum, not angular velocity. Centripetal acceleration falls as 1/r³, not the way Grace wrote it. For the actual dimensions of the ship the drop would be 50–60 times. To get “half a g” in the module after letting out the cables, as in the book, you’d have to spin it up to 25–30 g beforehand. Grace wouldn’t survive. The law of conservation of angular momentum left silently, without saying goodbye. I didn’t even notice it go.

Accelerating at Hover Altitude

After that the ship heads for orbit. At 2 g of thrust for ten minutes, nearly horizontally. At hover altitude.

Air density there is about three thousandths of a kilogram per cubic meter. It’s the upper atmosphere, but it’s atmosphere. Accelerating nearly horizontally to an orbital speed of about twelve kilometers per second at that altitude is atmospheric entry, just in reverse. Drag at 12 km/s gives about fourteen g. A ship with a twenty-by-ten-meter hole in its side becomes a meteor. A “stable orbit” with its periapsis inside the atmosphere is impossible in principle — such an orbit brakes itself within a few revolutions.

A Leak That Can’t Happen

Astrophage starts leaking out of the tanks through the hole in the hull. Fed Astrophage, let me remind you, reflects light according to the book. The engine beam hits the leaking paste. According to the book, the cells “absorb the momentum.” Fed cells can’t absorb, they reflect. So we get about 10¹⁵ watts inside the compartment, reflected off the paste in every direction. The ship would burn up entirely. On top of that, the heated ship starts glowing on the CO₂ lines, and the tanks see it. The layer of paste at the breach is about two trillion cells, on the order of 3·10¹⁸ joules, about seven hundred megatons. If it fires, everything else fires, and that’s 2·10²¹ joules. Emergency jettison of compartments won’t save you. The ship should have turned into a small supernova. Of course, it didn’t.

Ammonia

And the last thing. Rocky, saving Grace, breaks through into his compartment. The Eridian atmosphere breaks through with him: ammonia at 29 atmospheres and 210 °C. Grace survives.

The lethal concentration of ammonia for a human is half a percent to one percent. Equalizing 29 atmospheres with Grace’s compartment would give him fifteen to forty percent ammonia. Pulmonary edema, death within one to three days. But first — the shock wave. Equalizing pressures with a ratio of about 88 through a one-square-meter opening is a wave with a Mach number of about three, 3.6 bar behind the front, up to 18 bar reflected. Gas in body cavities gets compressed elevenfold. And for Rocky, conversely, pressure drops from 29 atmospheres to a third of an atmosphere, and his body, which lives at 210 °C, boils. Nobody survives. In the book: “Ammonia everywhere,” and both are alive.

Chapter 6. The Eridians, or A Civilization That Can’t Do What It Does

Rocky is the best thing in the book. Everyone agrees on that: critics, readers, and me. Even the dealmakers of the media industry agree: they react to the idea of selling five-legged spider merch the way Astrophage reacts to carbon dioxide. Only with an erection on top. A charming spider-like engineer who speaks in chords and ends every question with the word “question.” Except the longer I look at him, the more he scares me.

Radio Without a Transistor

The Eridians don’t have transistors. No vacuum tubes. No computers. The book repeats this several times: Rocky doesn’t know electronics, they don’t have light bulbs — in the epilogue Grace has to explain how to make them. They solve problems by “humming in a trance,” a hive mind.

And yet Rocky’s sampler has a radio transmitter in it. “A very standard Eridian data connection.” Without an active amplifying element — a transistor or a tube — a small radio transmitter is impossible. That’s not “backward technology.” That’s technology they don’t have but do have. They have a camera and a display — “old technology.” How does a blind species without electronics make a camera and a display? How does a civilization without computers calculate an interstellar trajectory? How does it build hull robots? How, finally, does Rocky figure out the Beetle’s computer controls in an hour if he “doesn’t understand electronics”?

Each of these points could have been closed with one sentence: “the Eridians use…”. The book closes none of them. It just says “they don’t have it” in one chapter and “here it is” in the next.

Space Is Empty

Rocky says: “There’s nothing in space. Space is empty.” The Eridians don’t know about cosmic radiation. Their crew — 23 people — died along the way, and they never figured out what killed them. Rocky survived because he slept next to the fuel tanks.

And yet the Eridians built a space elevator up to synchronous orbit. Erid’s synchronous orbit lies inside the planet’s radiation belt, and Erid’s magnetic field, according to the book, is 25 times stronger than Earth’s. Building an elevator for years inside a radiation belt and not knowing about radiation is like working in an X-ray room and not knowing about X-rays. They also have spectroscopy, they see the Petrova line, they know the atmospheric composition of a neighboring planet. They know chemistry well enough to synthesize xenonite. But the atomic nucleus — no, their atomic models have no neutrons at all. The number of beads equals the number of protons. With atomic masses twice the atomic number, chemists would have noticed that in the first week.

A Newtonian Plan That Flew by Einstein

The Eridians don’t know special relativity. So, according to the book, their ship took fuel with a huge surplus — “scads.” They planned to fly 6.64 years by Newton, but arrived in three — because by Einstein, time on board runs slower.

Let’s check. To live three years on board and cover ten light-years, the ship has to accelerate at about 2 g — which, incidentally, is exactly Erid’s familiar gravity. By Newton, in that time they’d have accelerated to three times the speed of light. A civilization that has radio and the Petrova line — that is, all of electrodynamics — calculates a flight assuming you can go faster than light? And nobody noticed that the speed of light in their Maxwell equations is a constant? Or the other way around, they calculated a slow Newtonian flight, and then “three years” doesn’t add up. Pick one. The book picks both.

The Fridge Tank

Inside the Blip-A: 29 atmospheres of ammonia at 210 °C. The Eridians live at that temperature. And what’s in the ship’s tanks? Astrophage. Which holds 96 °C and devours heat. A tank of Astrophage next to Rocky’s compartment is an inexhaustible refrigerator: it draws heat through the walls, and it can’t be “filled up.” What heats the ship? The book has no heating source. Rocky sits for 46 years next to a giant freezer and doesn’t freeze. At the end of the book, when his fuel is at zero, Rocky’s compartment should have cooled down within hours to a day. Six weeks later, Rocky is fixing tanks as if nothing happened.

And that’s only half the trouble. By canon, even fed Astrophage holds its 96 °C. Fed means it has nowhere left to charge. But heat from the 210 °C compartment keeps flowing through the tank wall, nonstop. Where to? There are no hungry cells in the tank — it’s a fuel tank. Into neutrinos? Then the fuel tops itself up, and after 46 years Rocky has more energy in his tanks than he had at launch — a perpetual motion machine with home delivery. Outside? Then it’s not a refrigerator, it’s a hole in the balance sheet that the book doesn’t show. Heat came in, went nowhere, the tank is warm, the fuel is all there, the compartment is warm too. In my youth, circumstances like these usually brought in the OBKhSS, the Soviet economic-crimes police. And its first question was always: where does the product go?

Worse: an Eridian body by itself is 210 °C, oxygen, carbon dioxide. That’s a perfect incubator for Astrophage. A gram of Astrophage getting into an Eridian draws up to nine megawatts of heat from the body. The lethal dose is micrograms. The Eridians’ only defense against disease, according to the book, is fever: infection gets burned out by temperature. And Astrophage feeds on heat. For the Eridians, Astrophage should be the plague, the most terrifying thing there is. And yet they breed it in the ocean in metal spheres and calmly haul it around in tanks on the other side of a wall from the living quarters. No quarantine. No infections. Wounded Rocky’s blood spills all over the Hail Mary, right next to the Astrophage, and nothing.

Their Own Farm

Where do the Eridians even get Astrophage from? To breed it in spheres, you need the first cells. Slow Astrophage only exists in the breeding layer at a planet with CO₂ or at the star itself. According to the book, the Eridians never left orbit before Rocky. You can’t catch a cell in the Petrova line at 0.92 c — a 20-picogram cell carries about three kilojoules and will punch through any collector. So either they flew to their own “Venus,” which isn’t in the book, or they know more than they’re telling. Remember this too.

A World Where Water Boils

A couple of words about Erid itself. Oceans at 210 °C. The saturated water vapor pressure at 210 °C is about 19 bar. If the atmosphere is 29 atmospheres, then two thirds of it should be water vapor, and water boils at 232 °C, so right on the edge. With a stellar flux of about 15.7 kilowatts per square meter — eleven and a half times Earth’s — the oceans would boil off unless the clouds reflect more than 92% of the light. This is a planet on the brink of a runaway greenhouse. Rocky, though, compares “29 times” to the air of the tunnel they built, which is at 400 hectopascals. Then the pressure on Erid is about 11.6 atmospheres, water boils at 187 °C, and oceans at 210 °C boil off. Again, pick one.

A day on Erid is 5.1 hours. With an orbit 0.2 astronomical units from the star, tides should have braked its rotation roughly thirty times faster than Mercury’s. You’d expect tidal locking or a resonance. No such luck. And there’s ammonia in the atmosphere. Ammonia is a “loaded spring”: the star’s ultraviolet would destroy 29 atmospheres of ammonia in ten thousand to a million years. You need a reducing source to replenish it. The book has nothing.

Rocky the Master Builder

And one last thing about Rocky. For 46 years he’s been hanging at Tau Ceti. His sampler broke off in flight, and he can’t make a new one. 46 years. An engineer who builds anything and everything.

And then he meets Grace and in a few days: builds a sphere to go out to the humans, builds a transfer tunnel, makes ten tanks with a nitrogen dosing system accurate to one part per million, makes copies of the tanks, melts “noncritical systems” into a “metal stew” and gets an alloy “stronger than any metal alloys” without refining. Xenonite, which “can’t be touched by any tool,” gets drilled with a power drill in one chapter, chipping “like glass,” and in another it’s “not brittle.” Rocky himself says of xenonite: “I not understand science. I just use.” The recipe: “liquid and liquid. Mix.”

The attentive reader will ask: isn’t it a bit too convenient how this engineer can do things one minute and can’t the next? I’ll ask that too. A bit later.

Chapter 7. Taumoeba, or The Piano in the Bushes

A Predator Turns Up

At Adrian, Grace finds Taumoeba — a predator that eats Astrophage. It has to be delivered to Venus. The problem: Venus’s atmosphere has nitrogen, 3.5%, and Taumoeba dies from nitrogen in seconds. In 23 generations Grace breeds the strain “Taumoeba-82.5,” which tolerates 8.25% nitrogen. Hooray.

First. Nitrogen is N₂, the most inert molecule in the atmosphere. It’s so inert that industry uses it when something needs to not oxidize. Meanwhile ammonia, NH₃ — caustic and reactive — is harmless to Taumoeba. The toxicity mechanism is the exact opposite of chemistry. And on Venus, at 3.5% nitrogen and 0.02 atmospheres in the layer, the partial pressure of nitrogen is about 70 pascals, and the nitrogen in a cell is on the order of five nanomoles. Taumoeba dies from five nanomoles of inert gas in seconds. Okay, fine. Let it.

Second. Grace calculates how much nitrogen should be in the tanks for the experiment: “Less than 1 Pascal.” Actually 3.5% of 0.02 atmospheres is about 71 pascals. An error of seventy times. In the head of the hero who “does relativistic calculations.”

Through Xenonite

Now the book’s main piano in the bushes rolls out. Taumoeba-82.5 starts passing through xenonite. The very same xenonite that no tool can touch. The one that nitrogen, a 0.36-nanometer molecule, can’t pass through. A micron-sized cell — passes. A channel a cell can squeeze through is, for a gas molecule, a highway more than a thousand times wider than itself. And also: if Taumoeba had eaten through xenonite chemically, it would have ruined the xenonite of all the tanks and of Rocky’s entire ship.

The selection was for nitrogen tolerance in tanks with a constant nitrogen fraction. A Taumoeba hiding in the wall starves and doesn’t reproduce. And a Taumoeba that has become tolerant has no reason to hide. These are two mutually exclusive traits: under constant nitrogen, selection for hiding loses. And passing through a centimeter of xenonite is a function Taumoeba never had at all. You can’t breed it from scratch in 20–30 rounds of selection. It’s like breeding concrete-drilling hamsters in twenty generations.

What does Taumoeba eat while it spends hours getting through a centimeter of xenonite? Destroying a centimeter-long channel takes 0.02–0.14 joules. The cell’s energy reserve is 10⁻⁶ to 10⁻⁵ joules. The deficit is a factor of a thousand to a hundred thousand. Either Taumoeba eats xenonite, or xenonite is metastable and collapses on its own, or this episode never happened. The book picks the third, only it first shows that it did happen.

Twenty-Two Million Kilograms Out of Nothing

Moving on. Rocky’s tanks hold “millions of kilograms” of Astrophage. Taumoeba ate it all. At the end Grace lives on Taumoeba from Rocky’s tanks, and there’s, according to the book, on the order of twenty-two million kilograms of it. But of millions of kilograms of Astrophage, the actual matter — biomass — is on the order of ten thousand to a hundred thousand kilograms. The rest is energy in neutrinos. You can’t grow twenty-two million kilograms of Taumoeba out of that: the gap is 200–2000 times. And without food, according to the book, Taumoeba dies within a week. The tanks hold dead mass.

And where, by the way, does the energy of the eaten Astrophage go? The 20 tons of fuel that “will turn into methane” are 1.8·10²¹ joules. If Taumoeba digests a cell, that energy has to go somewhere. The ship is intact only if all of it went into neutrinos. Convenient.

A Predator That Can’t

And the main thing. Taumoeba explains neither the stable equilibrium at Tau Ceti nor the Sun’s recovery. A predator with no dormant stage, no backup food, specialized on a single prey is a classic Lotka–Volterra system. It produces boom-and-bust cycles: the predator eats the prey, starves to death, the prey recovers, the predator returns. Tau Ceti should flicker with a period of about two years and an amplitude of up to 10%. It doesn’t flicker. And on Venus, where humans will seed Taumoeba, the predator will eat the prey and die out, and the Astrophage remnants on the Sun will rebuild the population. In 60–125 years the Sun will dim by 10% again. The book — with a happy ending.

Chapter 8. The Finale, or Robinson Teaches Friday

A Sacrifice Nobody Needs

The climax. Grace gets the Beetles and the fuel, he can fly home. But he learns Rocky is in trouble: Taumoeba ate his fuel. Grace turns around to save his friend and dooms himself never to return to Earth. Beautiful. Tears. Applause.

Now let’s do the math. Rocky’s ship is intact. He’s only out of fuel for braking at Erid — he can get there ballistically. How much fuel does braking take? With a dry mass on the order of three hundred tons and a speed of a few tenths of light speed — on the order of fifty to seventy tons of Astrophage. Grace has over two thousand tons. Three percent. Pour three percent into a steel tank — Rocky has steel tanks — and go your separate ways. The “top him up” option never comes up in the book. The sacrifice isn’t necessary. It’s only there to punch the reader in the tear ducts.

By the way, why did Rocky run out of fuel in the first place? A reserve of fifty to seventy tons, two to seven percent of the supply, could have been protected from Taumoeba by anything: nitrogen, steel, a separate compartment. Rocky had all of that. “All failed,” no explanation.

Flew to a Planet, Found a Guy

How does Grace find Rocky? Rocky has headed toward Erid, Grace sees the flash of his engine and turns his spin drives into a radar. He scans in planes every 5°, from 0 to 90°. He finds Rocky’s ship by its echo in 28 seconds. Honestly, this is the best idea in the book — a radar made from a photon engine, I like it.

But. Planes every 5° up to 90° cover half the sky. With a beam width of about a degree, 80% of it is gaps. The range window is about a million kilometers out of twenty. The chance of finding him is a few percent. That’s not luck, that’s “flew to a planet, found a guy.” The right way is to calculate Rocky’s trajectory, fly to intercept and scan along the way. The error in Rocky’s position is mostly along the track — about 10 million km, and hundreds of thousands across. A cigar, not half the sky.

And also: after a month of flight, light from Rocky takes over three days to reach Grace, Doppler shifts his 25.984 µm by 27%, and the exhaust points at 34° to the axis. Grace couldn’t have seen the flash of his engine every day, as in the book.

Sixteen Years Under a Dome

The epilogue. Grace is on Erid. He lives under a xenonite dome. His life support is maintained by 30 Eridians. He’s fifty-three and walks with a cane. He eats Taumoeba, meat grown from his own cells (“meburger”), and vitamins synthesized by the Eridians. In the first years he suffered from scurvy and beriberi.

Gravity on Erid is 2.1 g. A man over fifty at 2.1 g: blood pressure at the brain when standing up drops almost to the lower limit at which the brain maintains blood flow — fainting. Pressure in the foot veins is about 190 mmHg — edema and varicose veins. A fall from standing height is like falling from 3.7 meters on Earth, and for an older person that’s a broken hip. He should have wrecked himself in the first few months. A 1 g centrifuge on the Hail Mary in orbit, a pool to immerse himself up to the neck, a lying-down regime — the book has none of it.

Scurvy. According to the book, Grace had three months of regular food and forty months of “coma slurry” with vitamins — the coma nutrient mix. As a supplement to the calories from Taumoeba, that would have lasted fifteen to thirty years. Scurvy isn’t inevitable, it’s a price the author assigned. And also: Rocky says that proteins and sugars are shared between humans and Eridians. Heavy metals in Eridian food can be separated by ordinary chemistry — sulfides, chelates, ion exchange. Grace has a lab. Rocky has a food supply for 23 Eridians there and back. There’d have been no hunger. And no monodiet, which kills even lemurs.

The Sun Has Recovered

The “Astronomy hive” reports: the Sun has returned to full luminosity. The treatment took “less than one of your years.”

Light from the recovered Sun takes sixteen-plus years to reach Erid. The Beetles take almost thirteen to reach Earth. From Grace’s departure to the epilogue is about 27.4 years. Light from the recovered Sun would have arrived no earlier than 29.1 years plus the treatment time. The news arrives two to three years earlier than is physically possible.

And also. To say “full luminosity,” you need to tell 100% from 99%, that’s half-a-percent precision, stable over decades. The level of the Kepler and Gaia space telescopes. For a blind civilization without transistors, tubes and computers. With thermocouples and a galvanometer. The news can’t be verified. Remember this too.

Fifty Stars and Two Lives

Rocky and Grace reason: Tau Ceti has fewer than fifty neighbors, and two of them have life — so life in the Universe is common. The seeding was billions of years ago. Tau Ceti moves relative to the Sun at about 37 km/s — that’s 38 parsecs per million years. Four billion years ago the Sun and 40 Eridani were not its neighbors. The current sample says nothing about the frequency of life. And if the Astrophage ancestor spreads across the whole Galaxy, then once again we hit the contradiction of 98% mitochondrial identity. The book presents the heroes’ conclusion as correct.

Sixteen Years of Silence

The Eridians have the Hail Mary with its Petrovascope. Millions of kilograms of fuel. Wikipedia on a laptop. Grace, who knows how the Beetles work. A Beetle-class probe would get from Erid to Earth in seventeen and a half years, a beacon on the Petrova line in sixteen. The Eridians have a huge motive: Earth is the second species fighting Astrophage; there are infected stars all around; reinfection is inevitable. In sixteen years they don’t try to contact Earth even once. And Grace doesn’t even ask. The Eridians convene their biggest “thrum” for Grace’s life support. For contacting Earth — not a single one.

By the way, communication between Earth and the Hail Mary wasn’t thought through either. Grace himself, on one page, considers signaling with Petrova-line flashes and drops the idea. A single spin drive element — about nine terawatts, eight kilograms of Astrophage a day — would have given a channel of tens of bits per second. A report from Tau Ceti would have reached Earth a year before the Beetles. Pings, reports, news. Nothing.

What’s the Speed of Light?

And the last scene. Grace is teaching Eridian children. First question: “Who here can tell me the speed of light?”

A civilization that has an interstellar ship, a space elevator and a hive mind “much smarter than any Eridian” gets an Earth schoolteacher for the basics. For sixteen years they had Wikipedia and the Hail Mary. Their astronomers would have mastered special relativity in a year. But the Earthling teacher is needed by the plot, not by the Eridians. This is the “Robinson teaches Friday” scene. A scene that strokes the reader’s ego: “we Earthlings are better at something, at least.” This isn’t science fiction. It’s a nineteenth-century colonial novel in a spacesuit.

Chapter 9. What the Internet Says

When I’d collected a hundred and fifty items, I got curious: am I really the first? The book is five years old, it has two million ratings, a subreddit, a movie, articles in Scientific American. Surely everything has been picked apart down to the bone long ago.

I collected everything I could find: Reddit, StackExchange, Goodreads, blogs, popular-science articles for the movie release, reviews by scientists. It came to about seventy separate complaints from forty-odd sources. I checked them against my register. Here’s the result:

  • 30 items had already been found before me — and phrased roughly the same way.
  • 36 items were found partly: the topic was raised, but the conclusion was different or shallow.
  • 126 items I didn’t come across in the collected sample. An honest caveat: I didn’t read the Reddit comments in full, there’s a captcha there, so some of my “firsts” may be lying somewhere deep in the threads.
  • 13 items were found by others and missed by me.
  • Later the registry grew by another 88 items — a second pass through the chat plus two new holes. I checked those too: 12 had been found before me, 48 partly, 28 turned up nowhere, including a cell hovering at the Sun against light and gravity and Taumoeba hunting in gas.

What everyone found. The neutrino battery is a classic, people have been bashing it since 2021. Xenonite made from an inert gas. Mitochondria and panspermia — an eternal argument, and here astrobiologist Mike Wong backs me up: if everyone has mitochondria, then “Earth should be the source.” How the Eridians don’t know about radiation. How a blind species makes cameras. How Rocky reads a flat screen by echolocation. How all of humanity suddenly agrees to hand power to one person. How DuBois and Shapiro ended up in the same lab. A too-convenient coincidence in timing. The explosions in Antarctica. Neil deGrasse Tyson, on air, corrected Weir about a planet’s magnetic field — I have that one too.

What others found and I didn’t. Grace’s name on the mission patch and his face in the robots’ database, even though he was swapped in days before launch (Mary Robinette Kowal noticed this one). Mitochondria in a cell that lives in space with no oxygen and no electron acceptor. Tanks that, per the ship’s schematic, can’t hold two million kilograms. Eridians sleep only under supervision — so how did Rocky live alone for 46 years? The ship’s computer that doesn’t warn about closing in on an alien ship. Walking in an Orlan spacesuit at 1.4 g. A couple of minor “up-down” inconsistencies after the ship flips. Thanks, colleagues. Into the register it goes.

What I mostly found myself. Almost all the quantitative physics: braking at Venus, neutrons and tritium, the sterile breeding layer, the reflected exhaust at Adrian, links in the beam, the centrifuge and angular momentum, the spin drive train, Taumoeba’s energy in the xenonite channel. And almost all the mission architecture: Venus instead of the Sahara, one ship with no scouts, fuel with no safety catches, fifty thousand gigatons in low orbit, exhaust aimed at Earth. Apparently, for this you need to be specifically a thermal engineer with a nuclear past. An ordinary reader sees “spin drive” and thinks “cool.” I see “spin drive” and think “where’s the heat rejection.”

And about the “NASA scientists” who supposedly praised the science. NASA consulted on the movie, provided logos, arranged a screening on the ISS — that’s collaboration and PR, that’s normal. There’s no scientific review in a press release. The actual scientists journalists asked praise the internal consistency, the spin gravity, the relativity — and in the same breath call everything important speculation. I didn’t find any scathing breakdowns from scientists. Apparently scientists just don’t have the time. Well then, it’ll have to be an engineer.

Chapter 10. The Hypothesis That Explains Almost Everything

So we’ve gone through the whole book. A hundred and fifty holes. We could, of course, say the author just didn’t do the math. That’s boring. Let’s instead assume the author did the math, and everything we see isn’t mistakes. It’s evidence.

And then the picture comes together. Slowly, with a click, like a lock.

I have one rule, a detective’s rule: I don’t invent facts that aren’t in the book. I take only what Weir wrote himself — pages, numbers, quotes. And I put them together differently.

Part One: Sabotage on Earth

Who Makes the Decisions

Remember who makes the decisions in this whole story. One person. No council, no audit, with immunity “for any crime anywhere on Earth” and a pardon promised in advance. The decisions this person makes are systematically the worst possible. Not one or two. All of them.

Fuel is stockpiled on Earth — for years, on a quarter of the Sahara, even though there’s a free gas station right next door at Venus, which ArcLight has already visited. One ship. No backup, even though the ship was assembled from modules over sixteen launches and a second set could have followed it. A crew of three, with exactly one person per specialty. No scouts ahead. No communication with Earth: nobody on Earth will find out what happened to the mission for the next twenty-six years. No Plan B — no predator of our own, from Earth, for Astrophage. The Astrophage genome read and forgotten: “Analyzing its DNA. Good for them.” The line’s discoverer pushed aside. A fueled ship holding something like fifty thousand gigatons of energy hangs in low orbit over land.

One such mistake can be put down to haste. Two, to bureaucracy. When the mistakes march in a solid column and all face the same way, they’re no longer mistakes. It’s a project.

Baikonur

Three days before launch, an explosion at Baikonur kills both of the mission’s science specialists — the prime and the backup. For some reason they ended up in the same lab at a destructive test with a fuel that even the flight engineer didn’t know about. They were issued a milligram instead of a nanogram — a million times more. A speck visible to the naked eye, about twenty tons of TNT equivalent, instead of an invisible mote. Unlabeled. Stratt runs the investigation personally, in two days. Verdict: “dosage error.” Case closed, launch not postponed.

Any investigator starts with three questions: who benefits, who had the opportunity, and who controls the investigation. Here all three have one answer, and that person is the one running the investigation.

The Passenger

The vacated seat goes to a man who doesn’t want to go. A teacher whose only scientific thesis — “life without water is possible” — Astrophage has already refuted: it’s made of water. He hasn’t done biology in years. His qualification, per the book, was grown for years by Stratt herself: she pulled him into the project step by step, as if she knew in advance that a seat would open up. Meanwhile there’s a real volunteer, a chemist from Paraguay. Nobody trained him.

Grace is sent by force — “You’re a coward” — with his memory wiped. The drug is administered by the medical bed, into which a hidden program was loaded in advance, bypassing the commander. Commander Yao is lied to: he’s told Grace asked for the anesthesia himself. Yao, by the way, was against having people aboard against their will. Yao asked for a gun.

And here’s what I ask you to remember. This isn’t my fantasy, it’s canon: on Earth, the coma automation had already been secretly modified without the crew’s knowledge. The channel of interference is established by the text of the book itself.

Two of Three

Two of the three die in flight. Even though gene carriers, according to the study, survive the coma in one hundred percent of cases. Cause of death not established. No investigation. Bodies jettisoned without tissue samples, even though the cause of death decides whether Grace should go back into a coma for the trip home. Jettisoned by the one who remembers nothing and doesn’t know the procedures — that is, the only person who can’t notice anything.

The medical bed with a program secretly loaded for one crew member is the same automation that kept the other two in a coma. A drug that wipes one man’s memory is perfectly compatible with the other two never waking up at all. The book doesn’t rule it out. Quite the opposite: Ilyukhina has a kilogram of heroin aboard, DuBois a cylinder of nitrogen, and the crew selection has a box for “willing to go on a suicide mission.” Someone on Earth thought very carefully about how these people would die, and not at all about how they would survive.

Delay, Not Failure

Here the attentive reader objects: if you want the mission to fail, it’s easier to wreck the launch. Why all the complications?

Because the goal isn’t failure. The goal is delay. A mission that never launched means a scandal, a commission, an urgent Plan B. A mission in flight means hope. All alternative solutions get frozen with one formula: “we’re waiting for the Hail Mary results.”

And look how conveniently the observability is arranged. During acceleration the photon engine’s beam points back, at Earth: the ship’s Petrova line is visible for years, the mission is “alive,” everyone’s calm. Braking at Tau Ceti is done with the beam pointing away from Earth — invisible. Then silence until the Beetles, twenty-six years. A window of almost three decades — right at the peak of the disaster.

And the Beetles, which are supposed to bring the news? They’re known in advance not to make it. The ship has crew shielding — a millimeter of Astrophage, Lokken came up with it. The Beetles don’t, and the oncoming interstellar gas at 0.93 c burns out their cameras and electronics in minutes to a day. The vulnerability was known at launch. Nobody closed it.

And the passenger. An incompetent fool with amnesia isn’t an accident, it’s a guarantee. Literally every decision he makes on his own kills the ship: the burial in the exhaust, hovering over Adrian, the chain, links into the beam, accelerating in the atmosphere. Each time the plot forgives him until the next inevitable failure. Without Rocky it would never have got as far as the Beetles. A failure with no witness and no traces — and in twenty-six years everyone will say: “Well, we did everything we could.”

Who Needs This

Look at what exactly the mission is supposed to destroy. Not just “the Sun’s disease.” It’s supposed to completely destroy Astrophage — a free battery with the energy density of antimatter that charges itself from the Sun. A population worth a tenth of a percent of the Sun’s luminosity is twenty billion times all of the world’s energy production. An AA battery of Astrophage would power a house for tens of millions of years. If Astrophage stays, in twenty years Earth won’t need oil, gas or coal. Ever. Who loses out from free energy? Rhetorical question.

Not enough examples in history? Tetraethyl lead, sold as harmless for decades. Tobacco “research.” Exxon’s climate models, accurate inside the company and “uncertain” outside it. Gas-industry money for campaigns against coal. A public that for decades has been set against nuclear plants one day and wind turbines the next. Industries know how to defend themselves, and they do it not with explosions but with damage stretched out over time and responsibility smeared thin. Exactly like here.

And a cold world isn’t a catastrophe for everyone. In the cold, demand for fuel grows. Rich countries buy up oil and drop out of global competition. The hegemony that so irritates South Asia and North Africa sorts itself out — in favor of the oil states, India and China. The famine doesn’t hit those who organized it: the death of half of humanity is the death of anyone at all, just not the “golden billion” with the oil. Basically, ballast gets dumped. Hey, is there any downside at all?

Elite Panic

And a final touch. Stratt’s forecast of “food wars” and the death of half of humanity. How does she know? She doesn’t. It’s projection. Sociologists call it “elite panic”: the people at the top are sure that bestial chaos will break out below, and under that pretext they themselves behave worse than any mob. Rutger Bregman takes it apart in “Humankind” using the bombing of cities: the strategists were sure the bombing would cause riots and looting. It didn’t. People helped each other. And the strategists kept on bombing.

Stratt is the same breed. A sociopath in power who sees people as beasts because that’s how she herself is built. Her apocalypse scenario is the best way to get absolute power and immunity. She got it. And spent it on a mission that couldn’t come back.

Wait. Holy crap. That’s it.

Part Two: The Dark Forest in Space

Now let’s look at the Eridians. Through the eyes of someone who already knows he’s being lied to.

All the information about the Eridians in the book comes from Rocky. Grace never saw their planet. Never checked their history. Never examined the twenty-two bodies of the crew. Knows their biology from what he’s told. One witness, and an interested party at that.

A Backward Civilization That Can Do Everything

A civilization without transistors, tubes and computers. That has a radio data transmitter — “a very standard Eridian data connection” — a camera, a display, hull robots and an interstellar trajectory. An engineer who “doesn’t understand science, just uses it,” for 46 years “can’t” make a sampler, then in a few days makes everything the human needs, and in an hour bypasses a Beetle’s computer control, even though he “doesn’t understand electronics.”

A civilization that “doesn’t know” relativity, but arrived in exactly the time Einstein gives at their familiar gravity of two g. The 6.64-year “Newtonian plan” is a cover story. They flew by special relativity, at home gravity, and knew what they were doing.

A civilization that “doesn’t know” about radiation, but built an elevator through its own radiation belt. That “never left orbit,” but breeds Astrophage — and slow cells only exist at a star or at a planet with CO₂. So they’ve been out there. So they know more than they say.

All the Eridians’ “ignorance” works the same way: every time, it gives Grace the role of teacher. He explains relativity, radiation, computers, light bulbs. And every time he feels smart. For the reader it’s competence porn. For Grace it’s pleasure. In reality it’s taming sugar plus reconnaissance.

One-Way Exchange

Count who gave what. Grace gave relativity, radiation physics, the principles of electronics, human biology, Earth’s location, Taumoeba, Wikipedia and, finally, himself. Rocky gave finished products and labor: fuel, walls, repairs, xenonite as a product. Not one theory. About his main technology Rocky says: “I not understand science. I just use.” The recipe for xenonite: “liquid and liquid. Mix.” Thanks, very detailed.

And xenonite itself behaves in such a way that without a “secret ingredient” the physics and the chemistry don’t add up: one moment no tool can touch it, the next it’s drilled with a power drill, then it crumbles “like glass,” then it’s “not brittle.” A sample, if Earth ever gets one, doesn’t give technology. It creates dependence. Want more? Come to us.

And flattery from the very first visit: “You ship has more science than my ship. Better science.” Right after it, a request to move Rocky’s things onto the Hail Mary. Access to the whole ship obtained on the second date.

The Blind Man Who Hears Through Walls

Rocky, per the book, is blind and “sees” with passive sonar. Only sonar doesn’t explain what Rocky perceives. He makes out Grace’s fingers behind an aluminum plate and fine mechanisms, and for that you need megahertz. He hears Grace “two floors below” when Grace talks in a normal voice. Through one tunnel wall, speech is at the edge of audibility; through decks, no way. The Eridians’ nerves are light guides, and they have a camera: light is familiar to them both technically and biologically.

So Rocky most likely has a different sensory, and maybe signaling, system that he doesn’t show the stranger. And the scene where he “didn’t notice the lights went out” is blindness put on for show.

Mechanoid

Now read the description of Rocky’s body that the book gives in all honesty. Carapace: oxides. Bones: alloys. Blood: mercury, only his word for it, nobody ran a test. Nerves: silicate. Muscles: inorganic, steam-driven. Living tissue: “a few kilograms” out of 168. Two percent. A “hive” of cells services the structure. Sleep is paralysis while the hot loop gets serviced. That’s not sleep. That’s scheduled maintenance.

Guys, this is a mechanoid. Grown or manufactured, the text doesn’t let you tell. And the story about eggs, a mate and an age of 291 years is there so as not to scare you.

And inside him, most likely, is Astrophage. Per the book, Rocky already cools his suit with Astrophage: “Astrophage take all heat hotter than ninety-six degrees.” A body at 210 °C with a radiator is the ideal customer for such a refrigerator, and the stored mass is a battery. Three hundred watts over 46 years of waiting is about five milligrams of Astrophage. And three holes close at once: why Astrophage isn’t a plague for the Eridians (it’s not an infection but a domesticated organ), what heats the Blip-A, and why Rocky survived by the tanks when everyone else died.

Predator at the Top of the Food Chain

Rocky himself says the Eridians are at the top of the food chain. Predators. And a civilization of predators that had to domesticate plenty of living things just to avoid starving looks at everything that moves with one thought: “how do we domesticate this?” The only question is finding the comfortable regime. What suits a chicken won’t suit a horse.

Astrophage is their livestock. They “breed it in balls in the ocean,” and that’s a fairy tale that contradicts all of Astrophage’s biology: if it feeds on the ocean’s heat, migrating to the star is pointless, and an Eridian body is an incubator. So the fairy tale was told to Grace, and the herd grazes somewhere else.

Taumoeba is the sheepdog. Which suddenly “learned” to pass through xenonite, a material only the Eridians know how to make. Which gets bred in 23 generations for exactly the conditions needed — as if it wasn’t bred but switched on. And the “accident” with Taumoeba in Rocky’s tanks leads to exactly one outcome: Grace himself, voluntarily, flies to Erid.

And the mitochondria. Ninety-eight percent identity between Astrophage and Earth life. That’s not four billion years of separate evolution — humans and chimpanzees differ more than that. That’s recent contact. Someone, quite recently by cosmic standards, spread the same life to neighboring stars. For the Eridians, Earth is an ice planet. There’s no point in conquering it. Using it is another matter.

On Earth, colonization almost never started with genocide. It started with a trade mission and an outpost. With beads and little mirrors. With “we are your friends.” Humans as labor on ice planets the Eridians have no use for themselves: a great idea. The main thing is that the humans keep feeling like friends. For the time being.

Guard Post

And Rocky? Rocky waited at Tau Ceti for forty-six years. Not for help. For a guest.

Tau Ceti is the only uninfected star in the neighborhood. Anyone looking for a cure will come exactly here. The perfect place for an ambush. The “coincidence of eras” — two ships of two civilizations at one point in the same decade, odds of about one in ten million — stops being a coincidence. It’s not a meeting. It’s a post.

Fuel on the Blip-A: “scads,” a huge surplus. That’s not an error in a Newtonian calculation. That’s the reserve of a guard post that waits for years and must have the resources to maneuver. And, in one of the handling scenarios, to escort the alien to his home planet. In the book this literally happens: Rocky pumps more than two million kilograms of fuel into the Hail Mary — a full load for an interstellar trip. For himself he keeps “millions of kilograms.” The post refueled an alien starship.

And Rocky most likely isn’t alone there. Grace was simply “assigned” Rocky. Nobody knows about the others, and nobody will. Grace never got to see all of the Blip-A, and he knows about the twenty-two dead only from what he’s told.

Rocky found Grace at a point in space where finding anyone is impossible — as if it wasn’t Grace searching but Rocky.

Wikipedia

What does the post get? Grace hands the Eridians all of Wikipedia. Rocky asked for “Science information from Earth,” and Grace decided on his own to copy everything. That’s not “Earth science.” That’s the complete history of Earth: all the wars, all the genocides, all the colonizations, all the weapons, all the politics, all the vulnerabilities. And the lesson that follows from that history all by itself: a competitor must be outpaced, outgunned or destroyed. A handbook on “how our species works and where it hurts.”

The decision is made by one person who has no authority to speak for the planet. The Eridians get both the handbook and a live specimen. Grace has, in effect, sold out Earth lock, stock and barrel.

The Enclosure

And the specimen then lives like this. A dome of transparent xenonite — “because they thought that’s what I would want.” Thirty keepers, “or so I’m told.” Lights switched on and off on a 24-hour schedule. Meat grown from his own cells. Every day selected children are brought to him: “the selection process… Eridian culture is complicated.” According to the book, he’s “of extreme scientific interest.”

Guys, that’s an enclosure. That’s a zoo. Grace is a talking parrot that knows relativity. The Eridians, as we’ve established, know relativity themselves. They don’t need a lesson. They need a specimen of an Earthling in working order.

The news of the Sun’s recovery is brought to him by the “Astronomy hive.” It can’t be verified: half-a-percent photometry is impossible for a blind civilization without electronics, and the news itself arrives two to three years before the light could. Why report it? To keep the animal calm. In sixteen years the Eridians never once tried to contact Earth, even though they have a ship, fuel, a Petrovascope and a man who knows how the Beetles work. A probe would reach Earth in about seventeen and a half years, a beacon in sixteen. Either the communication goes on without him. Or they don’t want Earth to know where its man is.

The Dark Forest

The dark forest. A civilization that doesn’t know how its neighbors work doesn’t give itself away, but tames, studies and puts on a shelf. Two predator civilizations: one at our home, in an office with immunity from any crime, the other at a neighboring star, in a carapace of oxides. And between them, one kind, bewildered teacher who all this time was sure he was the savior.

The Earth half made sure nobody would miss Grace: the mission was set up so that nobody would learn anything about it. The Eridian half picked up what nobody is looking for. Perfect prey. Didn’t even have to negotiate with anyone.

And here’s the book’s last scene. A school classroom. Eridian younglings. An Earthling at the blackboard. “Who here can tell me the speed of light?” Robinson teaches Friday. The reader weeps with tenderness. And I see a cage, a sign reading “Homo sapiens, male, 53 years old, feed three times a day,” and a tour guide waiting for the parrot to finish its act.

Wait. Holy crap. That’s it!

Epilogue. Why This Matters

I’m not claiming Weir intended this. Of course not. Most likely he just didn’t do the math. But I like that the book, if you read it carefully, refutes itself so thoroughly that you can assemble a better story from the wreckage. Darker, smarter, more honest.

Good science fiction is a canvas. On it the author poses questions we don’t have answers to: what is a human, what is power, is it okay to coerce for the sake of salvation, whom can you trust when everything’s at stake. In “Project Hail Mary” all these questions are lying right on the surface. Absolute power without a counterweight. A man sent to his death against his will. A suicide crew. A gift from an alien civilization that could turn into a weapon. The book walks past every one of these questions with a smile and a “gosh” joke. “Her plan worked perfectly.” The end.

And then there are the rules of the genre. Pianos in the bushes are forbidden. If an author puts on a lab coat and says “neutrinos,” “mitochondria,” “ATP” and “0.92 c,” he takes on an obligation. This isn’t a scene from “The Big Bang Theory” where Penny slides her glasses down to the tip of her nose, languidly breathes “molecules” — and Leonard is in no state to check what came next. Although, judging by the raptures, Weir turns his readers on in roughly the same way. He’s telling the reader: “I checked.” Harrison doesn’t say that, and he can get away with anything. Asimov doesn’t say it, and he can get away with anything. Lem says the opposite — “we won’t understand” — and that’s great honesty. Weir says “I checked” on every page. And didn’t check.

When “The Martian” came out, we were thrilled: finally someone was writing SF where you can do the math. “Project Hail Mary” shows the flip side: you can adopt the “I’m a scientist” tone, the jokes, the formulas, and sell two million copies to people who won’t recalculate anything. That’s not fraud. That’s the market. And the hype, and the reader’s level, and the genre crisis. All of it together.

A Product, Not a Novel

And one last thing, for a sense of scale. MGM bought the film rights to “Project Hail Mary” in March 2020 — a year before the book came out, from the manuscript. The sum, according to Deadline, ran to seven figures. Same announcement: Ryan Gosling to star and produce, Lord and Miller to direct. Gosling read the manuscript at the start of the pandemic, when the book didn’t exist yet. So by the time the first reader opened the novel, it was already known whose face Grace would have on the poster.

Did Weir write with Gosling in mind? I don’t know, I wasn’t in the room, and there’s no direct evidence. But open the book again and look at how it’s built. It’s not a novel. It’s a set of scenes for the screen: short “problem — joke — solution” episodes, flashbacks sliced up one per crisis, visual attractions — the centrifuge, the chain over Adrian, the spider in the ball, the five-legged critter’s claw tapping on the wall. The characters’ inner life sits at the level of lines that are easy to deliver on camera. The holes we found are exactly where the scene needs things to look pretty and nobody feels like doing the math.

It’s a commercial project. In the style of “Harry Potter.” With unpleasant analogies: both in the quality of the prose and in the logic of the world, which Steinway & Sons use as the natural ecological reservoir for quality musical instruments. Well, you get it: wood. Very good, well-seasoned, resonant wood. Blockhead grade.

I did finish it, though. Rocky is nice. If you love him — love him. Just don’t tell me it’s scientific.

The second law of thermodynamics, by the way, has rejoined the chat. It asked whether it could please sit out the sequels from now on.


The full register is below. Holes (H): the book contradicts itself or the laws it itself acknowledges. Stretches (S): possible, but requires every character to switch off their brain at the same time. Numbers match the review log. The “Online” note at the end of each item shows whether it was found before me: “yes” — found, “partly” — the topic was raised, “no” — not encountered in the collected sample.

Appendix. Full Register of Holes and Stretches

Holes

H1. Direct side-on observation of a narrow Astrophage beam. A cell’s beam points strictly backward and cannot be seen from the side. The book honestly explains the visible arc by scattering on interplanetary dust, but that requires a brightness calculation that isn’t there. See H68 — a narrow beam from a single cell is impossible in the first place. Online: no.

H2. Exactly one observed frequency at different speeds. Cells on the arc accelerate, brake and fly at different angles to us. Because of the Doppler effect, the single narrow 25.984 µm line should come out as a smeared band. Online: no.

H3. Sun–Venus trip at 0.92 c. According to the book, ISRO measured 0.92 c for cells flying in both directions. Shedding that speed in ~300 km of atmosphere takes a deceleration of ~10¹⁰ g, which would smear the cell. Online: no.

H4. Photon rocket equation for the Hail Mary. With 2.1 million kg at launch and four years at 1.5 g, an ideal photon rocket loses mass by a factor of ~415. No more than 5 t reaches Tau Ceti, less than the weight of a single lab. Online: partly.

H5. Infection chain via WISE 0855−0714. The Sun supposedly caught the infection from WISE 0855−0714. That is a brown dwarf at ~250 K, invisible to amateurs, with nothing for Astrophage to feed on, and its “dimming” can’t be measured from amateur data. Online: partly.

H6. Aerobraking of Astrophage at 0.92 c (Venus and the Sun). A cell at 0.92 c brakes against the upper atmosphere of Venus (or the Sun’s chromosphere) over 20–300 km. That is 10¹⁰–2·10¹¹ g and 1–20 GPa, a pressure comparable to the strength of diamond. Even at a “slow” 0.01 c, atmospheric braking means about 4.5·10¹² J per kilogram of cell, and it will evaporate. “Absorbs any heat” doesn’t help: to stop, the cell must transfer momentum to a gas mass of ~γm, and in its frame each CO₂ molecule strikes with ~64 GeV; that means nuclear cascades, pions, gamma rays and neutrons, and almost all the kinetic energy stays in the gas. Online: no.

H7. Energy balance of Venus and the absence of a gamma glow. Even 0.01% of the power stolen from the Sun, arriving at Venus as kinetic energy, is ~10⁴ times Venus’s solar flux. The planet would overheat and glow in gamma rays, and Fermi-LAT would see it at once. Online: no.

H8. Energy for the return flight from Venus. To get back to the Sun at 0.92 c, a cell needs a reserve of ~3.9 mc². At Venus it would take tens of thousands of years to accumulate it. The return trip that ISRO “sees” is energetically impossible. Even recovering all the braking energy (1.55 mc²), the cell won’t collect 3.9 mc², let alone enough for two daughter cells. Taking into account the cell’s reserve after accelerating and braking at Venus, it is left with ≈35 dry masses against the ≈235 needed: the cycle doesn’t close. Online: no.

H9. Blackpanels in the Sahara. By the book, 1 m² of farm yields ~520 W in stored mass, more than the actual insolation of the Sahara. Besides, a cell at 96 °C is fed by foil heat at 120–150 °C, and Carnot allows only 6–13%: the farm is 8–15 times weaker than claimed. The blackpanel rate (0.5 µg/m² per day) is computed at 1000 W/m² around the clock, i.e. 520 W/m² on average, while the Sahara gives ~270 W/m². Besides, below 96.4 °C the output drops to zero, not by a few percent: night, winter, clouds and dust inside the panel, which can’t be washed out from inside, reduce it to nothing. Online: no.

H10. Proper time of the Beetles. At 0.93 c, over 12.8 years of Earth time, ~4.7 years pass aboard a Beetle. To get Hatch’s “20 months” you need ~0.9915 c. In the finale Grace also writes about “a couple of years” of the Beetles’ onboard time, so neither of the author’s two figures adds up, and they don’t agree with each other. Online: partly.

H11. Mass and density of an Astrophage cell. A 10 µm cell of 20 pg is 25 times lighter than water, and a charged one (17 ng) is denser than osmium. Neither looks like a living cell that is “mostly water,” as the book says. Grace fuels the Beetles with a syringe: 100 mL of paste weighs about 600 g, i.e. 6 g/cm³, denser than most rocks, and the book doesn’t comment on this. Online: partly.

H12. The Beetle’s cameras and electronics in the head-on flux. At 0.93 c the head-on gas delivers 7–8 Gy/s at a Beetle’s nose. Cameras and electronics die within minutes, even hardened electronics within a day, and the Beetle has no shield. There is nothing to shield the probe’s head with: 1.6 GeV protons require hundreds of grams of matter per cm² (a meter or two of water), and optical glass darkens already at 10²–10³ Gy, i.e. within minutes. Even at the real gas density of 0.1–0.2 atoms/cm³ the timescales grow only to days, and without accounting for nuclear cascades the estimate is still too low. Online: yes.

H13. Neutrino mass vs. cosmology. The book’s neutrino mass (~0.048 eV) gives a mass sum of 0.165–0.186 eV. Cosmology (Planck, DESI) rules that out. The number is fitted: the neutrino’s rest energy (≈0.0477 eV) coincides both with the quantum of the Petrova line and with 3/2·kT at 96.4 °C (≈0.0478 eV); DESI constrains the mass sum even more tightly, <0.07 eV. Online: partly.

H14. The neutrino battery. You cannot hold neutrinos in a 10 µm cell: their mean free path is light-years of lead. Annihilation νν → γγ is practically forbidden, and two quanta emitted in opposite directions give zero thrust. The Compton wavelength of such a neutrino (≈26 µm) is larger than the cell, and a particle locked inside it cannot be “at rest.” A cell’s reserve of ≈mc² is ≈2·10²³ neutrinos in 5·10⁻¹⁶ m³, and by the Pauli principle the Fermi energy of such a gas is ≈3–6 MeV, 10⁸ times the rest energy: annihilation would give MeV gamma rays, not 26 µm light. Online: yes.

H15. The error about a planet’s magnetic field. Grace explains Erid’s strong field by the planet being “in the star’s field.” A planetary dynamo is self-sustaining and needs no star: Jupiter’s field is the strongest, Venus has none at all. Proximity to the star harms the field rather than helps it: tides at 0.2 AU slow the rotation needed for the dynamo (S42), and the stellar wind compresses the magnetosphere. Online: yes.

H16. Erid’s ocean at 210 °C under a star 11.5 times brighter than the Sun — a climate on the brink of a “runaway greenhouse”. At 210 °C the water vapor pressure is ~19 bar, and two thirds of the atmosphere would be steam. With 11.5 times Earth’s flux the oceans would boil away unless the clouds reflect more than 92% of the light. At 29 atm water boils at about 232 °C, so an ocean at 210 °C has a margin of only ~20 degrees. With an ordinary albedo of 0.3 the planet absorbs about 2.7 kW/m², while a steam atmosphere cannot radiate more than ~300 W/m²; a continuous bright cloud layer would also explain the darkness at the surface, but the book has none. Online: no.

H17. Astrophage is “just dropped in the water” on Erid. By canon, Astrophage needs CO₂ and the light of its IR lines to breed. In Erid’s dark, alkaline, ammonia ocean there is no free CO₂ and no such signal. The Eridians put CO₂ in spheres, but a closed sphere has no IR lines, only blackbody light, and the second law doesn’t let you “feed” on uniform heat without a cold sink. Online: no.

H18. “Radiation doesn’t reach the surface” doesn’t make the Eridians radiation-naive. Erid’s atmosphere blocks cosmic rays, but potassium-40, uranium, thorium and radon give a background of the same order as on Earth. A species without DNA repair wouldn’t have arisen under such a background, and it would know radiation. Radioactivity can be discovered without sight: on Earth it was measured with an electroscope and an electrometer, and metallurgists smelting ores would have run into ionization of the air. Online: yes.

H19. The source of the Eridians’ first Astrophage cells is unexplained. Slow Astrophage occurs only at a star or at a planet with CO₂, and before Rocky the Eridians never left orbit. Where they got the first cells, and how they connected black dust with the star’s dimming, is not said. Online: partly.

H20. A 10 µm cell can’t aim at Venus. From the top of the arc Venus is seen at ~22″, while a 10 µm cell’s diffraction limit at 4.26 µm is ~24°. The cell would receive ~20 photons from Venus over the entire flight. It cannot aim at a planet. At 18.31 µm the cell’s pattern is even wider, ~105°, and refining the direction by brightness statistics is possible only with ~10⁷–10⁸ photons, whereas ~2.6·10⁴ arrive from Venus at 18.31 µm. Navigating by the magnetic field is not possible either: Venus has none, as Petrova herself notes. Online: partly.

H21. How ArcLight caught Astrophage from Venus orbit. ArcLight collects samples from Venus orbit over 12 passes. But above the atmosphere cells fly at ≳0.01 c, and a 20 pg cell at 0.92 c carries ~2.8 kJ and would punch through any collector. A soft catch is impossible. A cell is slower than ~10 km/s (the limit for soft capture in aerogel) only within ~0.5 m of its starting point or the end of braking, i.e. inside the breeding layer. Passing through this layer at an orbital ~7 km/s at an altitude of ~90 km is atmospheric entry in a single orbit, not 12 orbits. Online: no.

H22. A nanoneedle kills, a relativistic proton doesn’t. According to the book, Astrophage dies from a nanorod but not from relativistic protons. A proton stopping inside a cell breaks every bond along its track, the same mechanical trauma, only worse. At 0.92 c the dose is ~2.5·10⁷ Gy/s against ~10⁴ Gy lethal for microbes. For a cell at 0.92 c the solar wind is protons of ≈1.45 GeV, and during the throw to Venus (~7 minutes) it accumulates ≈4·10¹⁰–10¹² Gy: the cell would die on every trip within the Solar System. Online: no.

H23. “20 t of Astrophage will turn into methane”. Charged Astrophage is almost entirely stored energy: in 20 t of fuel there are only 25–650 kg of matter. The methane from it is ≲0.1 t, hundreds of times below Grace’s estimate. And where ~10²¹ J went is not said. When 20 t of fuel dies, ≈1.8·10²¹ J is released (≈430 Gt of TNT), and the ships survive only if all the energy went into neutrinos (M3): if even 1% turned into heat, the ship would evaporate. Online: yes.

H24. Taumoeba-82.5 passes through xenonite, but nitrogen doesn’t. If a cell micrometers across passes through xenonite, then a 0.36 nm nitrogen molecule passes all the more easily and thousands of times faster. And if Taumoeba gnaws xenonite, Rocky’s walls and tanks should degrade. Online: yes.

H25. Energy balance of the breeding layer and Taumoeba’s food supply. For Taumoeba to be fed, the layer needs thousands of cells per cm³. But that many Astrophage release heat and warm the layer to ~96 °C, and it rises as a plume. “Well-fed Taumoeba” and a “cold thin layer” differ by a factor of 100–10⁶. Online: no.

H26. Taumoeba explains neither the equilibrium at Tau Ceti nor the Sun’s recovery. A narrow predator without a dormant stage produces boom-and-bust cycles, hence a star flicker of up to 10% with a period of ~2 years. After Taumoeba dies out, the Sun would return to −10% within 60–125 years. Taumoeba explains neither the equilibrium at Tau Ceti nor the “cured” Sun. In a predator–prey model the prey level is set by the predator’s properties, so the near-zero level at Tau Ceti is a fudge, and a prey refuge at the star stabilizes coexistence but leaves the star dimmed. The ending requires the incompatible: the predator ate the prey down to zero and survived itself. Online: no.

H27. Astrophage needs any heat source >96 °C, not a star. By the book, Astrophage charges from any heat above 96 °C: Erid’s ocean, foil in the Sahara. Then it has no need to fly to the Sun; the optimal Astrophage would sit in Venus’s hot clouds, and the book’s premise wouldn’t exist. According to the Venus profile (VIRA), below ≈47 km the temperature is above the threshold (143 °C at 40 km), with 96.5% CO₂ and thermal glow in the 4.26 and 18.31 µm bands. Assembling a daughter cell costs ≈4·10⁻⁷ J against ≈1.5·10⁶ J of “fueling” for the trip to the Sun and back, so a sedentary strain would divide ≈4·10¹² times more cheaply, in minutes instead of years, and would displace the migratory one within days. Online: no.

H28. Heat balance of the Blip-A: Astrophage should have frozen Rocky. Astrophage holds 96 °C and absorbs all heat above that, while the Eridians need 210 °C. The tanks next to Rocky’s workshop are an inexhaustible megawatt-scale refrigerator, and the book has no heat source. For a ~1 cm wall at 210 °C inside not to be hot outside (p. 273), its thermal conductivity must be ≲0.005–0.01 W/(m·K); then the tanks draw off ~10 kW, which for a full crew weakens the hole to a stretch, but for Rocky alone over 46 years the deficit remains (~0.3 kW against ~10 kW). With such an insulator Grace would not have felt hot by the wall on p. 183; the text contradicts itself. Online: no.

H29. Astrophage should be a deadly infection for the Eridians. An Eridian’s body is 210 °C, with oxygen and CO₂, an ideal incubator for Astrophage. 1 g in the body drains ~9 MW, and a lethal dose is micrograms. The book has no quarantine and no infections. Online: no.

H30. The Eridians’ Newtonian flight plan and its outcome don’t add up. The Eridians’ Newtonian plan is 6.64 years, yet Rocky arrived in 3 with a huge leftover of fuel. By special relativity “3 years” corresponds to 2 g, but a civilization with radio and no special relativity is already a stretch, and the leftover fuel doesn’t match the plan. The Newtonian plan of “10 light-years in 6.64 years” requires ≈0.88 g and a maximum speed of ≈3 c, i.e. it is faster than light, while “3 years” under special relativity requires ≈0.996 c and ≈520–560 dry masses of fuel, more than the Newtonian budget (≈410–415). A Newtonian, moreover, would not have chosen the native ~2 g: by his model the mass ratio at 2.1 g is ≈11,000 against ≈415. Online: partly.

H31. Eridian radio without an active element. The Eridians have no transistors or tubes; Grace teaches them to make light bulbs. Yet Rocky’s sampler carries a “small radio transmitter” for sending data. A small transmitter is impossible without an amplifying element. Without tubes and transistors, the options left for a transmitter are the spark gap, the Poulsen arc and the Alexanderson alternator, all bulky and unsuitable for a small transmitter, and a crystal detector is already a semiconductor, a direct path to the transistor. Online: yes.

H32. Camera and display in a blind species without electronics. A blind species without transistors or tubes has a camera that sees light and a dynamic display as “old technology.” Earth’s pre-transistor analog is only tube television. Online: yes.

H33. The Eridians’ only defense is fever, and Astrophage feeds on heat. The Eridians have no immunity; they burn out infection with fever up to 305 °C. For Astrophage, heat is food. The wounded Rocky’s blood spilled across a ship full of Astrophage and nothing happened. Online: no.

H34. “Venus has no ionosphere”. Venus’s ionosphere has been known since Mariner 5 (1967), and there is also an induced magnetosphere. Only its own dynamo is missing. It doesn’t affect the plot, but an astronomer from Pulkovo wouldn’t have written it. Online: no.

H35. No funnel at the Sun: open field lines at the pole diverge. Open field lines at the Sun’s pole fan out. Over 37 million km the cells would form a cone tens of millions of km wide, not a thin line. There is also no southern arc and no polarity reversal every 11 years. Online: no.

H36. 98% identity and “four billion years” are incompatible. 98% mitochondrial identity means a recent common ancestor: human and chimpanzee mtDNA differ by ~9%. Grace’s “four billion years of divergence” contradicts his own figure. The common ancestor of Earth life is a prokaryote; mitochondria appeared late, through endosymbiosis. Earth life cannot descend from a “ready-made eukaryote” with mitochondria. Online: yes.

H37. “There are no other elements on the Sun” and phosphorus. “There are no other elements on the Sun except hydrogen” is false: the Sun holds millions of times more carbon than Venus’s entire atmosphere. And Adrian has no phosphorus or nitrogen for DNA and ATP. Online: partly.

H38. Hangar test: 1.8 GJ in 100 µs is an explosion, not a “muffled whump”. 1.8 GJ in 100 µs goes into 20–200 g of surface silicon. That is plasma and an explosion of ~0.4 t TNT on a deck with people on it, not a “muffled whump.” And that energy wouldn’t even melt a ton of silicon. Online: no.

H39. Polycarbonate is opaque in the mid-IR. The spin drive plates are polycarbonate, and polycarbonate is opaque in the mid-IR (4.26 and 18.31 µm): the stimulus won’t reach the cells. Even 10⁻⁶ absorption of 9 TW is 9 MW per plate, and it would melt. Online: no.

H40. Spin drive train: monolayer, oil, soot. Fed Astrophage reflects light, so only a monolayer can work. But even a monolayer reflects ~1/8 of the power back, the paste’s oil evaporates into the beam, and soot can’t be removed with a “scraper”: the gap is ~10⁷ times. The solution — a tape cassette with a monolayer and raster exposure — the user came up with in 10 minutes while walking the dog. The book doesn’t apply to the engine the properties of Astrophage that it introduced itself (to S82). A coherent layer without a mirror shines equally forward and backward, so a gold mirror is needed behind the cells, not transparent polycarbonate. A 1 µm oil film absorbing 0.1–1% of an element’s power receives 10–100 GW, and dead cells in the light of their neighbors char and fuse into the polycarbonate, which softens at about 147 °C. Online: no.

H41. “European languages had no word for a tornado” — factual error (doesn’t affect the plot). Europe gets ~300 tornadoes a year, and words for them have existed for centuries (trombe, tromba d’aria, waterspout). A factual error that doesn’t affect the plot. Online: no.

H42. “Earth has a negligible CO₂ signature” — error in the book (p. 91). Earth’s CO₂ bands are saturated, and as a CO₂ beacon Earth shines at 0.5–1.7 of Venus, not “negligibly.” By the book’s own logic a noticeable share of the flux would go to Earth. Online: no.

H43. A crew member’s body in the exhaust beam wrecks the ship by reflection. A body dropped under 1.5 g thrust drifts to the stern and into the beam. As it vaporizes, it reflects 10¹³–10¹⁴ W back, against an engine tolerance of ~10⁻⁶ of the power. A burial into the exhaust wrecks the ship, and there is no airlock interlock for thrust. Each of the nine fuel compartments has a Jettison button. An empty ~4 m tank in the beam is as much a mirror as the body, and a punctured tank with fuel is a bomb at the stern. The body’s plasma remains a mirror for ~1 ms, and then the cloud of vapor and debris, pushed by the light, scatters it back for up to seconds: even from 1 km the ship receives ~5 MW with a tolerance of ~0.5 MW per element. Amnesia doesn’t justify the ejection: DGSE doesn’t erase “trained skills” (p. 430), Grace on p. 141 remembers that the exhaust “would vaporize,” and the correct procedure is to cut the thrust, push the body sideways and restart the engines once it leaves the beam cone. Online: no.

H44. “29 times the atmosphere” — relative to the Hail Mary’s air, not Earth’s. Rocky compares “29 times more” with the tunnel’s air (400 hPa); he has seen no other human air. Then Erid is at ~11.6 atm, water boils at ~187 °C, and the oceans at 210 °C boil away. Online: partly.

H45. Hovering at Adrian: reflected exhaust hits the engines, the 30° tilt doesn’t help. While hovering at Adrian, reflected and scattered exhaust sends 10¹¹–3·10¹² W back against a tolerance of ~6·10⁸ W. The spin drives would die within seconds to minutes, yet in the book only a hull plate melts through. A 30° tilt doesn’t help: plasma reflects across the beam. Online: partly.

H46. The chain and the thrust vector: 30°, ≈100 m/s and 2° are incompatible. A chain hanging under thrust always trails toward the exhaust side and lies flatter than the beam. At 30° and ~100 m/s it is almost horizontal (82–88°), and at 2°, as in the book, it hangs straight in the beam. For the ship to hold altitude and move at constant speed, the drag of the whole system must be ≈0.577 Mg; with low drag (the chain deflected by only 2°) the ship accelerates sideways at ≈g·tan30°, and “a little over 100 m/s” is only an instantaneous value. Online: no.

H47. Solar particles confused with the solar wind. 400 km/s is the solar wind (~1 keV), which is not radiation. The danger is flare protons at 0.15–0.9 c, and you can’t outrun them in 8 hours. An error for a physics teacher and a shielding designer. Online: no.

H48. Braking at Venus glows with neutrons and makes Astrophage radioactive. Braking cells at 0.92 c against Venus’s atmosphere means hadronic showers: ~10³² neutrons/s, ~10 mSv/h at the ISS, 4·10⁵ times more C-14 in Venus’s CO₂. The word “neutron” never appears in the book. At 1% dimming this is ~1 Sv/h on the ISS, a lethal dose within hours, and spacecraft neutron detectors would see Venus as the brightest neutron source in the sky. 1 mg of cells grown from C-14-enriched CO₂ would give ~100 Bq; any beta counter would point to nuclear reactions at Venus. Online: no.

H49. Tritium in the cell sterilizes the breeding layer. Showers from braking cells produce tritium. In a water-filled cell it gives ~0.1 Gy/s, and at ~50 km above Venus ~3·10³ Gy/s, so the breeding layer is sterile. ArcLight’s samples would glow with tritium. Tritium in DNA or protein turns into helium-3 and breaks the bond from inside; the canon “the cell quenches external particles” doesn’t help against this, and Deinococcus grows only up to ~60 Gy/h against ~10⁴ Gy per day in the cell. 1 mg of cells contains no less than 10⁸ Bq of tritium, and any lab’s liquid scintillator would see it at once. Online: no.

H50. Centrifuge: Grace’s law is wrong (p. 357). The coil’s internal motors don’t change angular momentum: when the tethers are paid out, g-load falls as 1/r³, by a factor of 50–60, not 14. To end up at “half a g” afterward, you need 25–30 g before payout. Online: no.

H51. Accelerating to orbit at hover altitude is atmospheric entry (pp. 348–349). The ship accelerates “almost horizontally” at hover altitude, where the density equals that at ~40 km on Earth. At 12 km/s that means ~14 g of drag and Apollo-class heating. A stable orbit with such a perigee doesn’t exist. At Adrian the density at hovering altitude is ~3·10⁻³ kg/m³, and with “almost horizontal” acceleration the drag at just 4 km/s is ~1.6 g: the 2 g thrust goes almost entirely into drag long before orbital speed. First a steep climb of another 40–60 km is needed, not “a little above the horizon.” Online: no.

H52. Leak: fed Astrophage can’t “absorb the momentum” (pp. 351–352). Fed Astrophage reflects light, it doesn’t “absorb momentum.” 10¹⁵ W inside a punctured tank would be reflected around the compartment and burn the ship. The layer at the breach is ~700 Mt and would ignite all the fuel. A thrust of 6 g, ~2·10¹⁵ W, evaporates a ship of ~100 t within milliseconds, and the red-hot walls glow at 4.26 and 18.31 µm and become “Venus” for the cells in all compartments: all 20 t discharge (~400 Gt of TNT). The ship would explode before reaching orbit and jettisoning the compartments, and a jettisoned compartment with a breach is the same kind of bomb at the stern. Online: no.

H53. The Sun is the brightest “Venus,” and the magnetic stage is mandatory one minute, optional the next. By the book, a cell’s thrust follows the brightness of CO₂ in the IR, and at 4.26 µm the Sun is ~4·10¹¹ times brighter than Venus, so the cell would fly into the Sun. The magnetic launch is mandatory in one place, yet unneeded in Grace’s experiments and in the spin drive. Light pushes a starving cell ~3 times harder than gravity pulls (it blows away like dust), while a well-fed one needs ~1.4 W to hover with an inflow of ~5 mW — its whole reserve in ~2 weeks. Neither can “hang out on the surface.” The only consistent scheme is stages (charging → start along the field → search for CO₂ → division → flight toward the bright): ArcLight’s cells were caught at the search stage, so experiments without a magnet work, but the fuel from the farms was collected at the charging stage and without a field would not have switched on either in the spin drive or in the breached tank at Adrian. And if the transition is possible without a field, Redell’s magnet is unnecessary, and a cell at the Sun immediately sees its IR: again a flight into the Sun. Online: partly.

H54. Dropped links fly into the beam. A released chain link falls against the thrust, that is, into the beam. Each of the ~200 thousand links returns ~10¹² W to the stern, like ~25 kg of TNT at the nozzles. This is H43 repeated 200 thousand times. Online: no.

H55. Grace’s arithmetic: “less than 1 Pascal” (p. 451). 3.5% of 0.02 atm is ~71 Pa, not “less than 1 Pa.” An error of 70 times. “A hundred doses” would then weigh ~21 kg across seven tanks, more than the whole cylinder (10 kg). Online: no.

H56. The Hail Mary’s air composition doesn’t add up (pp. 122, 295, 394, 479–481, 486). The compartment is at 0.33 atm, of which oxygen is 0.2 atm; the remaining ~0.13 atm is unnamed, and nitrogen is aboard “only thanks to DuBois.” If the remainder is nitrogen, the ship’s own air would have killed Taumoeba-82.5 at 80 times the limit. If 0.33 atm is pure oxygen, as on Apollo, the phrase “one-fifth atmosphere” (p. 295) is wrong and the fire risk next to the burner and hot ammonia isn’t discussed; if the rest is nitrogen, as on Skylab, DuBois’s cylinder is unnecessary, and slides “in sealed bins full of normal air” (p. 486) are sterile without an experiment. Online: yes.

H57. Xenonite: its machinability contradicts itself (pp. 161, 275, 325, 370–371, 474, 486). No tool can work xenonite, yet an Earth drill bores it with chipping “like glass,” and at the same time it is “not brittle.” Chipping “like glass” is brittleness. “None of our tools would be able to cut it” (p. 474) is wrong even by the book: xenonite crumbs from the drill (pp. 370–371) in a resin binder cut xenonite, as diamond powder cuts diamond. How the Eridians themselves make molds, tolerances and holes for fasteners is not said. Online: partly.

H58. Rocky “couldn’t” protect his fuel, though he had the means (pp. 479–480, 494, 509). Rocky has nitrogen in any quantity, Taumoeba doesn’t pass steel tanks, and the compartments buy hours of delay. “All failed” with no explanation: there were means to protect at least the 50–70 t needed for braking. After 4–5 weeks of acceleration the Blip-A was already moving at ~0.15–0.2 c on a course to Erid (~50–65 years of flight with an Eridian lifespan of ~689 years), and the only trouble was overshooting the target: the braking reserve is ~16–22% of dry mass, 2–7% of “millions of kilograms.” Hiding in the wall doesn’t save Taumoeba from nitrogen: to eat the fuel it has to come out into the paste. Online: no.

H59. What keeps Rocky warm for 6+ weeks without fuel. Rocky’s compartment at 210 °C in interstellar darkness loses megawatts, and only Astrophage can heat it. Without fuel Rocky would have cooled within hours to a day. The text claims both “no fuel” and “life support works.” The compartment’s heat loss over 6 weeks is ≈2·10¹³ J, only ≈0.2 g of Astrophage: if grams for heating survived, tons for braking could have survived just as well. Online: no.

H60. The selection regime couldn’t have bred “hiding in xenonite”. Selection ran at constant nitrogen. A Taumoeba that hid in the wall starves there, doesn’t reproduce and doesn’t become a “champion.” Selection couldn’t fix “hiding” and “tolerating nitrogen” at the same time, and passing through xenonite is a function from scratch. A refuge also weakens the selection pressure for resistance; that is exactly why refuges are deliberately left with Bt crops. Two independent new traits, as with combination antibiotic therapy, require the product of the probabilities and don’t appear within ~20–30 rounds of selection. Online: yes.

H61. What Taumoeba pays to pass through 1 cm of xenonite. Gnawing a 1 cm channel through xenonite costs ~0.02–0.14 J, while a cell’s entire reserve is 10⁻⁶–10⁻⁵ J, and there is no food inside the wall. One of three things remains: Taumoeba eats xenonite, xenonite is metastable, or the episode can’t happen. With a xenonite density of 2–4 g/cm³, a 1 cm channel is ≈2.4 µg of material, ~5000 cell masses, and even 1 mm of wall costs 10²–10⁴ times more than the cell’s reserve, so Grace’s experiment with the disks (passage “in a couple of hours”) doesn’t add up either. If xenonite releases energy when it breaks down, it is an explosive, not armor. Online: partly.

H62. 22 million kg of Taumoeba from “millions of kilograms” of Astrophage (pp. 493, 511). Biomass can’t exceed the matter eaten, and charged fuel is only 0.1–3% matter. “Millions of kilograms” yields ≲10⁵ kg of Taumoeba, not 22 million. And without food it dies within a week, so the tanks hold dead mass. Online: no.

H63. The “radar” search — “flew to a planet, found a guy”. A plane rotated by 90° covers only half the sky. A 5° step with a ~1° beam leaves ~80% of directions unlit, and the range window is ~1 in 20 million km. The chance of finding Rocky is a few percent, yet he is found. Widening the beam to close the gaps isn’t possible: the reflection weakens about 25 times. At ~20 million km ordinary radio works, and Grace could have called Rocky by radio, and Rocky could have set up a beacon, which Grace himself thinks about (p. 497). Online: no.

H64. Daily tracking of Rocky’s flash is impossible without corrections the book doesn’t have. After a month, light from Rocky takes ~3.4 days, Doppler shifts 25.984 µm by ~33 µm, and the Petrovascope is tuned to the exact frequency. Without corrections Grace couldn’t have seen Rocky’s flash daily. Online: no.

H65. “Fewer than 50 stars near Tau Ceti, two with life” — the stars have mixed over billions of years. The seeding happened billions of years ago, and Tau Ceti moves at ~37 km/s, ~38 pc per million years. The current neighborhood is accidental, and “the 50 nearest stars, two with life” says nothing about how common life is. Moreover, Tau Ceti is older than the Sun, metal-poor and orbits the center of the Galaxy on a different orbit: 4 billion years ago the Sun and 40 Eridani were not its neighbors. Online: no.

H66. Scurvy and beriberi preventable with Grace’s own supplies. Grace has 3 months of food and 40 months of “comatose” mixture with vitamins. As a supplement to Taumoeba, that covers 15–30 years. Scurvy and beriberi were not inevitable. Years of a monodiet also mean depression (“even lemurs don’t live on bananas alone”), and the stock of Earth food allowed for a more varied diet. Scurvy kills within months, wet beriberi within weeks to months, and Grace could not have survived to the epilogue without traces of these vitamins in Taumoeba. On the other hand, ~10 mg of vitamin C per day is enough against scurvy and ~1 mg of B₁ against beriberi, so a supply stretched 5–10 times covered years. Online: partly.

H67. An Astrophage cycle with ≈100% efficiency violates the second law. The cell takes the Sun’s thermal light (≈5800 K), stores it in mass and gives it back as an almost monochromatic directed beam. The limit for such a cycle at 369 K is ≈91.5%, so ≈8.5% must go off as heat, and the book has no heat removal. On top of that, a 10 µm cell at 96.4 °C can radiate ≈0.33 µW, while near the photosphere it absorbs ≈5 mW, 15,000 times more: it cannot hold a constant temperature. Online: partly.

H68. A cell smaller than its wavelength cannot emit a narrow beam. A ≈10 µm cell is smaller than its 26 µm wavelength and radiates broadly, like a dipole: the divergence is more than 2 rad. The claim “each cell shines strictly backward” is false in itself, regardless of whether the beam can be seen from the side (H1). Only a coherent layer of ≥10¹² synchronized cells would give a narrow beam, and the book has nothing of the kind. Online: partly.

H69. Well-fed Astrophage next to a warm compartment: where does the heat go. By canon, even well-fed, fully charged Astrophage holds 96 °C. Heat from Rocky’s compartment (210 °C) flows into the tank without stopping, and the well-fed cells have nowhere to put it. If into neutrinos, the fuel recharges itself (a perpetual motion machine); if outward, it needs a sink that isn’t there. Complements H28. Online: partly.

H70. A cell can neither hover at the Sun nor charge there. At the photosphere a cell absorbs about 6 mW, while hovering against the Sun’s gravity takes about 40 mW: holding position eats more than comes in, and with that cost included the cell never charges. Even without this cost, the reserve for accelerating to 0.01 c (~450 J) takes ~20 h to collect, to 0.1 c (~4.7 kJ) ~9 days (twice as long with braking), the full mass-energy of a dry cell ~87 days, and the reserve for an interstellar trip at 0.9 c with braking ~4.3 years (~5.5 years at 0.92 c). A cell can store energy only in free fall or on ballistic loops, and how much it charges before departure the book doesn’t say (see H53). Online: no.

H71. A cell has no time to brake with light before Venus. From the Sun to Venus is about 108 million km, which at 0.92 c is 6.5 minutes. With its own thrust (~4 m/s² at 6 mW) the braking distance from 0.92 c is about 3.7 light-years, and the cell would fly past Venus long before stopping. With the same thrust, the “accelerate half the way, brake half the way” scheme gives at most ~660 km/s and a trip of ~3.8 days; so 0.92 c cannot be the cruising speed either with atmospheric braking (H3, H6) or with light braking, which moreover doubles the energy reserve needed for the trip. Online: no.

H72. The cells’ braking light would burn Venus. If the cells brake with light rather than against the atmosphere, each one shines forward, straight into the planet. With a flux of ~4·10¹⁹ cells/s (~2·10⁷ kg/s) the braking power is 10²²–10²³ W, while the entire solar insolation of Venus is ~3·10¹⁷ W. Even the minimal estimate of ~10²⁰ W at 0.01 c would heat the atmosphere to 1000–4000 K, and it would start to escape: changing the braking method doesn’t remove the H7 catastrophe. Online: no.

H73. A cell cannot lead its aim on Venus. The throw from the top of the arc lasts ~7 minutes, light from Venus takes another ~6 minutes to reach the cell, and in that time Venus moves by ≈28,000 km; aiming “at the visible point” misses by ~2.3 planetary diameters. To hit, the cell must know Venus’s angular velocity (~3·10⁻⁷ rad/s) and calculate the lead with an accuracy of ~12″ with its own angular resolution of about a radian, against the background of the Sun, which is ~10¹⁴ times brighter than Venus. All that remains is continuous homing along a pursuit curve, but that requires seeing the target the whole time and runs into the same diffraction limit (H20). Online: no.

H74. At 0.92 c a cell loses Venus because of Doppler. At 0.92 c everything ahead is blueshifted by a factor of ~4.9: the CO₂ lines at 4.26 and 18.31 µm arrive at ≈0.87 and ≈3.74 µm, and at its “native” wavelengths the cell already sees a different part of Venus’s spectrum (originally ~21 and 90 µm). One’s own speed cannot be measured from inside free flight without spectroscopy of external lines, and aberration also shifts the apparent directions. The cell cannot correct its course in flight; all that remains is a blind throw calculated from ~20 photons. Online: no.

H75. The spin drive irradiates its own ship: a cell must shine forward 10³–10⁵ times more strongly than backward. An incoherent layer at best shines like a Lambertian surface: 25% of the energy goes out at more than 60° to the axis, about 3% almost along the hull; that is ~4·10¹⁴ W onto protruding structures at ~9·10¹⁵ W at launch, and if half the light goes backward through a transparent plate, ~7·10¹⁵ W, about 1.7 Mt of TNT per second. The ship’s radiators (~10⁴ m² at 1500 K) would dissipate ~3·10⁻⁷ of the power, and a gold mirror absorbs ~0.5% at 26 µm, so a cell can give off backward no more than ~6·10⁻⁵ of its power, and windows in front of the cells are impossible (sapphire transmits IR only up to 5–6 µm, diamond would need an absorption fraction of ~10⁻¹²). The drive works only if the cell itself produces a narrow beam, and it can’t (H68); a coherent layer, on the other hand, is a weapon: with a 1 m nozzle the spot at 1 AU is ~4000 km and ~700 W/m². Online: partly.

H76. The canonical 4 s discharge rate overheats the substrate. Even in the best design (a monolayer ribbon on diamond, a hungry neighbor ~10 µm away taking the heat), with a cell discharging over 400 s the spot under it overheats by about 100 K, and at the canonical rate of 4 s by about 10⁴ K. So the cells must be discharged ~100 times more slowly than in the book; then ~2.6·10⁹ cells must glow simultaneously on each element, and a raster can’t be drawn with a scanning beam; a moving projection mask is needed. The book doesn’t discuss the “4 s” limit as a weak point (see H40). Online: no.

H77. The flat tunnel partition can’t hold 28 atm. The partition in the tunnel is a “big, un-reinforced flat panel” “half an inch” thick. A flat wall of radius ~0.75 m and thickness ~1.3 cm under a 28 atm pressure difference carries bending stresses of ~7 GPa, higher than any known material; even at ~11 atm (the H44 variant) it is ~3 GPa, the level of Kevlar and carbon fiber. The canonical 29 atm make the wall impossible even for a “magic” material (see S56). Online: partly.

H78. The death of the Blip-A crew from cosmic rays is incompatible with life at 210 °C. At 210 °C DNA is damaged by heat orders of magnitude faster than by radiation, so the Eridians must have extremely powerful repair, and with it radiation resistance: the hyperthermophile Thermococcus gammatolerans tolerates ~30 kGy, Deinococcus 5–15 kGy, while 4–5 Gy kill a human. Unshielded galactic rays give ~0.3–0.7 Sv/year; for a human that means cancer years later, not mass cell death, let alone the death of 22 out of 23. The book gives almost no symptoms or timelines (“Everyone get sick. Then everyone die”), so Grace diagnoses “radiation sickness” without grounds (see S90). Online: partly.

H79. The Eridian sphere in the ocean is incompatible with a staged life cycle. The Eridians breed Astrophage in spheres in the ocean (p. 327): the cells charge and divide without a magnetic start and without flying to a star and to a CO₂ target. If the stages of the cycle are optional, Astrophage has no need at all to fly to the Sun (H27); if they are mandatory, the sphere doesn’t work (see H17, H53). Online: partly.

H80. Adrian’s sticky box can’t catch the cells. At Adrian, samples are taken with a box of sticky resin exposed for 21 hours. If a layer of ~10⁹ cells settles on a ~0.1 m² panel, those arriving at 0.92 c carry ≈3·10¹² J (~0.7 kt of TNT), and those departing, accelerated to at least 0.01 c, ~10⁸ J. Indestructible cells would punch through the resin and the hull rather than stick (see H21). Online: yes.

H81. A 10 km chain doesn’t reach the breeding layer. With the thrust tilted at 30°, a ship of ≈120 t and a chain of 2–10 t, the chain is deflected 82–88° from the vertical and hangs down only 0.3–1.4 km. For the chain to give the required ≈1 MN of drag at 100 m/s, the air density must be ~0.5 kg/m³, otherwise the ship simply accelerates sideways. The chain and the beam can be separated only by a weight as massive as the ship (thrust and beam power double) or by a separate sampler (see H46, S124). Online: partly.

H82. Astrophage in the breeding layer can’t hover: hovering costs 800 times more than the inflow from the star. Hovering on photon thrust costs P = m·g·c: ≥9·10⁻⁵ W for a dry 20 pg cell and 0.075 W for a charged 17 ng one. Together with heat loss at 96.4 °C into gas at −51 °C this is ≥1.5·10⁻⁴ W per cell, 800 times more than it receives from Tau Ceti (≈2·10⁻⁷ W); a charged cell spends ~0.42% of its mass per day and can hover for ≈4.4 years. For the layer’s heating not to exceed the stellar heating, the cell density must be ≤0.02 cm⁻³, and then Taumoeba meets prey once in months or years (see H25). Online: no.

H83. Taumoeba has nothing to keep it in the layer: no buoyancy, no dormant stage. Taumoeba is ~2·10⁴ times heavier than the layer’s gas (0.02 atm, −51 °C), and holding up 1 ng takes a hydrogen bubble of ≈0.3 mm; that is no longer an amoeba. The scheme “it divides faster than it falls, and updrafts lift it” doesn’t help: turbulence would smear Taumoeba over kilometers of altitude, and almost all the time it would be outside the 200 m layer where the food is, and in downdrafts it would fall even faster. All models of life in clouds (Sagan and Salpeter for Jupiter, Seager et al. for Venus) require a dormant stage, and Taumoeba has none; nor can it hover inside clusters of Astrophage on their engines, because it kills a cluster within seconds and itself shuts down its own “elevator” (see S53). Online: no.

H84. Rocky’s exhaust points away from Grace. After ~30 days Grace (1.5 g) is moving at ≈0.13 c ~340 AU from Tau Ceti, Rocky (2.1 g) at ≈0.18 c ~460 AU away, their courses diverge by ~95°, and Rocky’s exhaust is directed at ~34° to Grace’s line of sight. With a narrow beam only the side lobes are visible, ~10⁻¹⁹ W/m², and without correcting for light travel time the error in Rocky’s position is ~100 AU. Either the beam is wide (and then Grace’s “radar” weakens, H63), or there was no daily tracking (see H64). Online: no.

H85. Near the Sun a cell must hover now against light, now against gravity, and the “balance point” is a charge level, not a place. In the book cells “hang out on the surface of the sun gathering energy via heat”. But sunlight and solar gravity both fall off as 1/r², so their ratio does not depend on height: there is no equilibrium point in space, only an equilibrium cell mass. For a 10 µm cell at the photosphere (intake ≈5 mW) that is ≈60 pg. A hungry 20 pg cell is lighter: light pushes it 3 times harder than gravity pulls, so it has to thrust towards the Sun all the time, shining its exhaust outwards. After ~2 weeks, at 60 pg, it passes the balance point and from then on must thrust away from the Sun, shining into it. The balance itself is neutral: there is no restoring force, and every absorbed joule makes the cell heavier and moves it off balance, so holding position always costs energy — |m·g − F_light|·c. While hovering the cell can store energy only up to ≈120 pg (twice the balance mass), ≈0.6% of the canonical 17 ng charge; beyond that, hovering eats the whole intake. Over the ~5 weeks to that limit ≈30% of the absorbed energy goes back into the Sun as exhaust and ≈16% goes outwards in the 25.984 µm line. At 10% dimming that is ≈1.6% of the Sun’s luminosity in a single line, ~170 times more than the Sun emits at all wavelengths longer than 20 µm: the whole star would glow in the Petrova line, not just an arc to Venus, and it would have been found at once. The solar wind hardly matters here (its pressure near the corona is ~1% of light pressure). An orbit does not help either: a hungry cell cannot stay in orbit at all, and a cell that keeps getting heavier sees its orbit shrink, which together with Poynting–Robertson drag (weeks near the Sun) takes it into the photosphere. Complements H70 (which covers hovering against gravity for a 0.55 ng water cell) and H53. Online: no.

H86. How Taumoeba catches Astrophage in gas. In the book Taumoeba is a single-celled predator, an “amoeba”, but how it finds and engulfs prey in gas at 0.02 atm and −51 °C is never said. Aquatic amoebae have pseudopods and something to crawl on; ciliates and flagellates have cilia and flagella; all of that works in liquid. In thin gas (mean free path ≈1.7 µm, comparable to the cell) you cannot crawl, swim or sweep prey towards you: cilia give a hundred times less thrust there than in water, less than the cell’s weight, and Astrophage on its own engine is thousands of times more mobile. What is left is a lottery: Taumoeba falls at 1–4 cm/s through hovering Astrophage and eats whatever it bumps into. That way it combs 0.001–0.05 cm³ of gas per hour, and to stay fed it needs tens to thousands of cells per cm³ — so many that the layer overheats (H25, H82). A predator that relies on luck would look different: a large capture surface, like the tentacles of a hydra or sea anemone or the silk of ballooning spiders. A sticky net of ~1 mm filaments would raise the cross-section 10²–10³ times and bring the required prey density close to the layer’s thermal limit, but the book’s Taumoeba is a bare cell. Moreover, the exposed membrane that phagocytosis needs dries out within seconds in Adrian’s dry gas (CO₂, methane, argon, no water listed), and at −51 °C the cytoplasm of a water-based cell freezes, so pseudopods have nothing to move with; a protective coat like a spore’s would make engulfing prey impossible. How Taumoeba eats, the book does not explain (see H25, H83, S53). Online: no.

Stretches

S1. Spin drive: coupling of the cells to the ship. The cell receives the photon momentum. How it transfers the force to the plate and where the heat goes, the book doesn’t explain. Online: no.

S2. Power and strength of the spin drive. 1.5 g on 2.1 million kg is ~9 PW of directed light. The force across 1009 elements is manageable, but even small absorption in the plates requires cooling that isn’t there. For a Beetle weighing a few pounds, with a beam of ~1.8·10¹⁴ W (~42 kt of TNT per second), removing waste heat is even harder than for the ship. For a 1000 t ship at 1.5 g the beam is 4.4 PW, and even an absorption fraction of 10⁻⁶ (4.4 GW) requires ~3000 km² of ideal radiator at 400 K. If the mandatory ~8.5% cycle losses (H67) were released on the ship, that would be ~375 TW. Online: no.

S3. Detecting the line from Earth, especially from Pulkovo. At 25.984 µm the atmosphere and a warm telescope give an enormous background; you need a dry, high-altitude site and a special instrument. Pulkovo, at sea level near St. Petersburg, is the worst place for the discovery. Faint mid-IR sources are isolated by subtracting an adjacent patch of sky, but for a wide extended arc the adjacent patch also contains its glow, and background subtraction eats the signal itself. Online: no.

S4. Solar glare. The top of the arc is ~14° from the Sun, so it can be observed only by day or at twilight, against the glare. Online: no.

S5. Xenonite. Strength, hardness, transparency and casting from two liquids at once require unknown chemistry. Weir himself calls xenonite “magic stuff.” The molds for the links are aluminum, and the Eridians bond xenonite firmly to aluminum — how is the casting separated from the mold? With xenonite’s super-insulation, the curing heat can’t be removed from thick castings. There are no movable sealed xenonite joints in the book, only fixed bonds and rigid chain links, and the material’s elasticity, fracture toughness and density are not specified. Online: yes.

S6. Scattering of the Petrova line on dust. The arc’s visibility rests on dust scattering light, but the dust density, the scattering efficiency at 26 µm and the arc’s brightness are not calculated. With an initial dimming of 0.01% and a fraction scattered toward Earth of 10⁻¹¹–10⁻⁹, the arc’s brightness is ~0.26–26 MJy/sr, while the zodiacal dust background at 25 µm, even 90° from the Sun, is 16–81 MJy/sr (COBE/DIRBE). There is no strict prohibition, but the arc is at best comparable to the background. Online: no.

S7. Interstellar medium at γ ≈ 10. At γ ≈ 10 each proton of the head-on gas carries ~8.6 GeV. 1 mm of paste won’t stop such a flux without Astrophage “magic,” and the hull in front of it produces secondary neutrons. At γ ≈ 10 and 1 atom/cm³ the nose receives ~400 kW/m², about 300 solar constants, and how the nose removes this heat is not said. Online: yes.

S8. The Beetles and the interstellar medium. The Beetles have a head with cameras in front and no protection against head-on gas at all, although the Hail Mary has an Astrophage layer. Online: yes.

S9. Tether centrifuge. Oscillations of the tethered assembly barely damp and require active damping, while in the book “the computer does everything.” The risks are resonance, jerk and breakage. The tether’s transverse mode is about 4 Hz, the longitudinal mode about 0.9 Hz (felt as a pulsation of weight), and the swinging of the compartment and the linkage has a period of 10–20 s. Running on the treadmill, pumps and footsteps, if they hit the natural frequency or its double, will rock the system until it breaks. Online: partly.

S10. Coma survival. In primate carriers of the gene, coma survival is 100%, yet in flight 1 of 3 survived. The book doesn’t discuss this. Online: yes.

S11. Stellar navigation of the Beetles at 0.93 c. At 0.93 c aberration squeezes the sky into a ~22° cone, and Doppler shifts the colors. An ordinary star tracker won’t recognize the constellations; a catalog recalculation is needed, and not a word about it. Ahead, with a Doppler factor of ×5.25, the camera sees what stars emit at 2.1–3.7 µm (an IR catalog is needed), behind, at ×0.19, the 76–133 nm range, where stars hardly shine, and undistorted colors are visible only on a ring ~47° from the course. The forward camera dies under the head-on gas (H12), so only the side view remains, and this is not discussed. Online: no.

S12. Handling a substance at 96.415 °C. The paste holds 96 °C with “infinite heat capacity” and can’t be cooled. A hull with such a layer must be insulated from the crew, and any low-temperature operations with the fuel are impossible. Online: no.

S13. Astrophage as heater and radiator. Astrophage is used as a heater or refrigerator only in isolated cases. At 9·10¹⁶ J per kilogram, that is strange modesty. Online: no.

S14. Sahara farms and heating Earth with Astrophage. A quarter of the Sahara under glass is 6–23 Gt of glass against a world output of ~0.1 Gt/year. Nobody proposes heating Earth with Astrophage from Venus (~4300 t/year for 10%). Besides the glass, ~2·10¹² static magnets and IR filters are needed, and just installing 2·10¹² panels would take ~1.5 years even with 10 million workers at one panel per minute. The book doesn’t give the construction time for the farms. Online: partly.

S15. Operating the Sahara farms. Dust, repair and harvesting across 2 million km² mean 1–2 billion cleaning robots, 10¹⁰ panel replacements a year and millions of people. In the book the farm “regularly delivered what was promised.” Training 1–3 million specialists (5–30 million with manual labor) takes longer than the whole project, and they would have to be taken from the glass industry, electronics and logistics, which make the panels themselves. Add 10 Sahara states, camps for millions, the meningitis belt and insurgents. The panels can’t be washed with water: at ~0.2 L/m² that is ~12 km³ per year, about a seventh of the Nile’s flow, and dust eats 0.3–1% of the light per day and requires cleaning every 1–2 weeks. Replacing ~10¹⁰ panels per year consumes almost the entire world output of sheet glass, and sand also abrades the glass, reducing transparency by percents per year. Online: no.

S16. The missed “Astrophage economy”. The flux at Venus is ~2.5 million t of Astrophage per day; the Hail Mary’s fuel passes by in 1–2 minutes. A population at 0.1% of the Sun’s luminosity would yield 20 billion times the world’s energy supply. It is destroyed without being considered. By canon, Astrophage absorbs any radiation and particles without heating above 96 °C — ready-made armor and a reusable storage medium for pulsed nuclear sources. The agent first rejected the idea, then admitted the mistake. Hatch says in passing “breeding the stuff on the moon or something,” but no one picks up or evaluates the idea. Online: partly.

S17. The 10% plateau and Taumoeba stability. An identical 10% for all stars is a fudge. Taumoeba at Venus, without a dormant stage, reserve food or repeated reseeding, would give boom-and-bust, not a permanent victory. A shortage of CO₂ can’t explain the plateau: Venus has ~4.8·10²⁰ kg of CO₂, while a population at 10% of luminosity needs on the order of 10¹⁵ kg. A mechanism is needed that hits a dense population harder and scales with the star’s luminosity, and there is none. Online: no.

S18. Eridian nuclear physics. Ignorance of radiation contradicts the Eridians working in radiation belts. Ignorance of the nucleus is implausible for chemists: iron’s atomic mass of 56 at charge 26 immediately points to neutrons. Without radiogenic heat the geology of the interior doesn’t add up, without thermonuclear fusion neither does the energy of their own star, and calculating the thrust of Astrophage engines runs straight into E = mc². On Earth radioactivity was discovered by chemistry, the electroscope and calorimetry, without sight. Online: yes.

S19. Erid’s space elevator without a rocket prehistory. Erid’s synchronous orbit is ~17,300 km above the surface, Δv ~19 km/s. An elevator can’t be built without a rocket program, and the book has none. Online: no.

S20. Eridians: oblateness, calendar, “there was no need,” and the Astrophage source. From the top of the elevator to escape velocity is only ~4.4 km/s. “There was no need” to fly beyond orbit is weak, and how the Eridians noticed the star’s disease is not shown. How the Eridians noticed the dimming (the climate responds over decades; it is visible at once only from orbit) and connected it with Astrophage is not shown. Rotation is noticed without light only through daily pressure variations and tides, and the year is almost invisible from the surface, so the Eridians obtained precise astronomy only from orbit, and then their ignorance of radiation is even stranger. The star (≈15.7 kW/m²) provides almost all of the surface’s 210 °C, and a 10% loss lowers the equilibrium temperature by 11–13 K: from the surface this is a trend over decades, from orbit it is immediate. Online: no.

S21. Eridians: “space is empty,” planets and electromagnetic radiation. Rocky considers space “empty,” although the Eridians work in radiation belts. They know spectroscopy and the Petrova line without understanding radiation, and they process all of it without a computer. The Eridians know that Astrophage flies through the “emptiness” to Thirdia, they study their strong magnetosphere and know the composition of Thirdia’s atmosphere, which is impossible without spectroscopy. The accuracy of Rocky’s astrometry (~10⁻⁵ for Adrian’s mass and orbit) requires telescopic detectors, and what instruments a blind species uses is not said. Online: yes.

S22. Mars is ignored. Mars’s CO₂ signal is ~60 times weaker than Venus’s, but in a chaotic search some cells would still head for Mars. The book simply doesn’t notice Mars. The CO₂ column above 1 m²: Venus ~10⁶ kg, Mars ~160, Thirdia up to ~400, and Thirdia, with a pressure <1% of Earth’s, is a working breeding planet, so “Mars is too cold and thin” is no argument. Online: no.

S23. Chronology of the epilogue (tentative). Light from the restored Sun would reach Erid no earlier than ~29 years after the Beetles’ launch, yet the epilogue takes place 1–3 years earlier. Online: no.

S24. The dimming exponential and the “earliest stage” (2.1). With a pure exponential, at the moment of detection the loss is already ~0.27%, three times the solar cycle. That is not the “earliest stage.” Online: no.

S25. Early weather anomalies (2.1). Tornadoes and frosts at −1…−2% are plausible, snow in Vietnam is overdoing it. The book gets the scale of the threat at −10% right. The assessment “snow in Vietnam is too much” should be withdrawn: in the book the hurricane season shortens, which is consistent with cooling, and cooling lowers the snow line and makes regional snowfall possible. Online: no.

S26. Methane from Antarctica (2.1). Compensation would need 400–1400 Gt of methane, while the ice that gives +1 cm of sea level holds ~180 t. A blast at 50 m throws up soil and precipitation, “minimal radiation” is wrong, and 20% of the bombs’ energy is neutrons. 245 (241 in the original) charges at 50 m mean ejected soil and a radioactive cloud (underground containment needs ~560 m for 100 kt), and 20% of the energy goes into 14 MeV neutrons; the ice will cool the ocean by ~0.0002 K, not by a degree. Subglacial carbon lies not in the ice but in sediments beneath the glacier (hundreds of Gt) and comes out over centuries, not in a month, and floating shelf ice barely raises sea level when it melts. Gas hydrates on the seafloor make more sense, but their methane will mostly dissolve and oxidize in the water. Online: yes.

S27. Epilogue and climate (2.1). Luminosity returns within a year, but the climate takes decades, and the ocean and glaciers take centuries. The epilogue is silent about this. Returning luminosity from 90% to 100% adds ~24 W/m², while the released greenhouse gas remains (excess methane decays with a characteristic time of ~12 years), and after the Sun is “cured” Earth gets a second catastrophe: overheating. Winding down the measures must be coordinated with the Sun’s recovery, and the book doesn’t notice this. Online: no.

S28. Unnoticed dimming of neighboring stars. Sirius and α Centauri are photometric standards (Hipparcos, Gaia, TESS). A 10% dimming would be noticed at once, and the infection chain can’t be reconstructed from a few years of amateur data. Online: no.

S30. Coincidence in timing. Two civilizations at neighboring stars reach interstellar flight in the same decade. Given a history of life spanning billions of years, the probability is ~10⁻⁷, with no explanation. Online: yes.

S31. Eridian biochemistry and energetics. DNA breaks down at 210 °C (the record for Earth life is ~122 °C). A high-energy species at the top of a dark food chain is not justified. Online: no.

S32. The chain scene at Adrian. A 10 km chain of 200 thousand links is cast by hand, although the Eridians have a xenonite cable for the elevator, and drill rods or a cable would be more economical. Plus photon thrust in the atmosphere creates a hot spot on the scale of a nuclear explosion. With a photon exhaust, every 1 MN of thrust is ~300 TW of radiation absorbed in the atmosphere: a continuously fed fireball with shock waves and X-rays right in the layer from which a live sample is taken. Online: no.

S33. Acoustics of the partition and passive sonar. Sound through the xenonite partition crosses two boundaries with a huge impedance mismatch, yet Grace hears Rocky. A passive sonar that distinguishes fingers behind aluminum requires megahertz frequencies. Bending stiffness grows as the cube of thickness, and a strong, rigidly fixed plate rated for 28 atm works poorly as a broadband membrane: the sound should already be loud on the human side. A vibrating wall in a random sound background destroys the spatial structure of the sound, and fine details require frequencies of ~100 kHz–1 MHz (wavelengths of 3.4–0.34 mm), which are absent from an ordinary background. Online: yes.

S34. Eridian physiology. Mercury blood can’t carry water-soluble substances without envelopes, and the internal O₂↔CO₂ cycle yields no energy by itself. Growth inside a carapace is not described. Darkness at the bottom of an atmosphere doesn’t rule out vision (deep-sea fish with eyes and bioluminescence), and sources online point out that Erid gets ~7 times as much light as Earth. Online: partly.

S35. Eridian knowledge, records and society. The writing, archives and specifications of an engineering civilization are not shown. One Rocky’s memory is no substitute for them. Online: partly.

S36. Builders of a relativistic ship who don’t know special relativity. Maxwell’s electrodynamics leads to Lorentz, and tests of a near-light-speed drive would reveal the discrepancies at once. Builders of a relativistic ship without special relativity are implausible. Online: partly.

S37. Politics and mission design. Stratt gains absolute international power without resistance from states. The only specialist is sent by force, with no backup missions and no second target. Lokken is taken with the phrase “You just got drafted” without ESA’s consent, Leclerc works under pressure, and people who are “very annoyed with each other” are put in one group. The team’s reliability rests on Stratt alone, a “Swiss watch with a single gear.” Online: yes.

S38. Energetics of Erid’s biosphere. Erid’s surface is “pitch dark,” yet the biosphere receives “an incredible amount of energy.” Geothermal flux is 10⁴ times less than the stellar flux, and the heat of a hot ocean is not food. Online: partly.

S39. Tau Ceti is the obvious suspect for the source. Tau Ceti sits in the center of the infected cluster, next to where the trail is lost, and shows no symptoms. This is a classic natural reservoir, yet nobody on Earth voiced the hypothesis. Of the stars in the chain, only ε Eridani (5.5 light-years) is closer than 8 light-years (the book’s range limit) to Tau Ceti, and the trail breaks off exactly at the only star that Tau Ceti can infect directly; 40 Eridani was most likely infected by ε Eridani (6.3 light-years). Online: partly.

S39a. Tau Ceti could have been sitting on the 10% plateau for a long time. Tau Ceti could have dimmed by 10% long before photometry began and been on the plateau ever since. This is hard to tell from Earth, and nobody tried. Online: no.

S40. Earth doesn’t pose the scientific question of the mission’s target. The alternatives “not infected / contained / long on the plateau” are not discussed and not tested by observations that are incomparably cheaper than a ship. The decision is one ship, no return, to an unverified target. What Earth will do with the answer is not thought through: the Beetles carry only data (5 TB), no samples, and Grace himself wonders whether to “throw Astrophage into Tau Ceti.” The average gas density along the route would have been cheapest to measure from Earth via absorption lines in Tau Ceti’s spectrum, but even that wasn’t done. Online: no.

S41. The Sahara project is infeasible given its combined resources. The number was reserved for a summary entry and never filled. It needs 6–23 Gt of glass (hundreds of years of world output), plus 1–2 billion robots, 10–35 million personnel (100–300 million with manual labor), and 1–30 million specialists. Fuel accumulates in 4.7 years by the book’s rate, ~9 years with the real Sun, 40–70 years with the Carnot limit, and 1000–1800 years with the fuel mass corrected per H4. Brings together H9, S14, S15, S77, S85. Online: partly.

S42. Erid’s 5.1 h day at a 0.2 AU orbit — tides should have braked its rotation. At 0.2 AU tides brake Erid ~30 times faster than Mercury. Tidal locking or a resonance would be expected, not a 5.1 h day. Such a day can’t be explained artificially either: spinning up the planet requires ≈1.6·10³² J. Online: no.

S43. Erid’s ammonia atmosphere: source and stability — a “loaded spring”. Ammonia is a “loaded spring”: ultraviolet would destroy 29 atm in 10⁴–10⁶ years. A source that replenishes it is needed, and the book has none. The oxidation 4NH₃ + 3O₂ → 2N₂ + 6H₂O gives ~330 kJ per mole, and the Eridians’ own oxygen biosphere would consume such an atmosphere, as in the Great Oxidation Event. The replenishment needed is ~3·10¹⁴–3·10¹⁶ kg/year, i.e. 10¹⁴–10¹⁶ W: within reach of an industrial Astrophage civilization, but not of a species without the transistor. Online: no.

S45. Shape of the Petrova line: sharp angle at the apex. Turning the arc by ~109° at 0.92 c requires a stop at the apex and a mass ratio of ~576. A smooth rise with a toss doesn’t explain why the whole arc is visible. Online: no.

S46. Grace’s lab experiments and the thermal background. In a “dark” cabinet the walls at 20 °C glow at 4.26 and 18.31 µm brighter than Venus, and the lamp glass is opaque beyond ~5 µm. Grace’s experiments show not what he infers from them. A bright lab and a dark cabinet differ only in visible light, to which, by the same experiments, the cells don’t react, while in the IR the conditions are identical. The only justification could be the halogen lamp of a microscope illuminator (quartz transmits up to ~4 µm), but the book has none. Online: no.

S47. The way back to the Sun. After division both cells fly toward the Sun, and the backward flux is twice the forward one. The Petrova line should be a Sun–Venus–Sun “triangle,” yet the book has one arc. A cheap return by falling (shedding ≈35 km/s at Venus and ≈620 km/s at the photosphere, 400–7000 times cheaper than accelerating back) would remove the hole, but the book closes it off: ISRO measures 0.92 c for cells flying away from Venus (p. 87). Online: no.

S48. Other organisms in the Petrova line at Adrian. Only organisms with their own engine (new photon rockets) or “passengers” on Astrophage can reach Adrian’s orbit. How bacteria and amoebae get there is not said. Online: no.

S49. Breeding layer at Adrian <200 m thick. The scale height there is ~3 km, and over 200 m the CO₂ pressure changes by 5–8%. A layer thinner than 200 m can’t be set by pressure alone; a sharp boundary is needed, and there is none. Cells departing toward the star fire their exhaust downward, and the dense CO₂ atmosphere absorbs 26 µm light, so why “IR clouds” are visible from above is unclear. The patchiness, on the other hand, means that reproduction depends on local conditions and not only on CO₂ pressure. Online: no.

S50. “Jupiter-like” bands and Adrian’s rotation. Many zonal bands “like Jupiter’s” are a sign of fast rotation, while tides at 0.54 AU should have braked the rotation. Bands and tides are incompatible. Online: no.

S51. 1:1 flux balance and mortality. An exact 1:1 between arrival and departure means Astrophage is immortal in transit. But the book itself says it dies. Online: no.

S52. Nitrogen as a poison. N₂ is the most inert form of nitrogen, while reactive ammonia is harmless to Taumoeba. The toxicity mechanism runs opposite to chemistry, and death “in seconds” at ~70 Pa is unexplained. Online: partly.

S53. How Taumoeba stays in the breeding layer. Taumoeba settles at 1–4 cm/s and crosses a 200 m layer in 1.5–6 h, with no buoyancy. How it stays in the layer and how it would take hold at Venus is not explained. Online: no.

S54. A “Plan B” isn’t discussed on Earth. Fast selection (23 Taumoeba generations), unlimited Astrophage and control of it on the farms are all available, but there is no parallel work on Earth on a predator of their own. Online: no.

S55. Interstellar “sporulation” is unprofitable for Astrophage itself. A cell that flies off to another star drops out of the population. Selection would remove such a strategy within a few generations; the book rests on group selection (“It spores!”). Online: no.

S56. A flat hull despite free-form casting. Rocky casts xenonite in any shape, yet the Eridian ship is made of flat faces. A flat 5 m slab at 29 atm is 64 times heavier than a sphere, costing an extra ~2600 t of fuel. Online: no.

S57. “The Eridians didn’t invent the spacesuit”. A spacesuit is needed not for vision but for pressure and temperature, in the elevator and in thin layers. The “no sensory input” argument misses the point. Online: no.

S58. Xenonite not sent to Earth. Rocky gave the xenonite away, but all of it stayed on the Hail Mary. Not even a few grams of sample were put on the Beetles. Online: no.

S59. Why Rocky waits at Tau Ceti for 46 years. Waiting 46 years without a sampler is pointless. He could have gone home and sent a second expedition, and Erid’s term of “≥72 years” has no starting point. The Blip-A’s full thrust, gathered into a ~1° beam, would be orders of magnitude brighter than the star in the narrow band of the Petrova line, and the Eridians already watch that line. A signal “alive, sampler lost” would have reached Erid in about 10 years, and a second expedition would have launched decades earlier. Online: no.

S60. Interstellar communication by Petrova light. Grace himself considers communication by Petrova-line flashes, and in the epilogue calls it impossible. A beacon would cost kilograms of Astrophage per day, and Earth–ship communication isn’t thought through either. A 1 TW transmitter (~1 kg of Astrophage per day) gives ~1 bit/s at a JWST-sized receiver, the Hail Mary’s full power 10⁴–10⁵ bit/s, and Earth has a receiver (MIRI works up to 28 µm). A chain of N relays raises the throughput on a segment N² times, so the limit here is not physics but the 32-year delay. Online: partly.

S61. “Several thousand words in a few hours”. 3000 words in ~5 hours is one word every 6 seconds, with frequencies entered by hand and the “gavagai” problem. Hundreds of words in the first day would be plausible. Online: yes.

S62. Earth’s science for Erid on a single laptop. All of Wikipedia on one laptop is ~45–50 years of reading through electronics that live 10–20 years. A single point of failure where several machines could have been handed over. Online: no.

S63. Wikipedia as a “scientific” gift. Wikipedia is not science but a tertiary source with errors and propaganda. It gives the Eridians the history of wars, dark-forest logic and knowledge of Earth. One person, Grace, made the decision. An Astrophage ship at ~0.9 c is itself a relativistic projectile, and the Eridians learn the history of wars, that humans have Astrophage, and Earth’s location: the “dark forest” works both ways. From Wikipedia the Eridians will verify physics and chemistry by experiment, but not history and politics. Online: no.

S64. Grace’s age (3.1). By the book’s timeline Grace is ~29 at launch and 23–26 at recruitment. He couldn’t have fit in a PhD, a scandalous paper, leaving science and a teaching career: a 5–10 year discrepancy. Online: no.

S65. Jack of all trades. A biologist-teacher alone does relativistic calculations, spacewalks, programming, chemistry and repairs. The workload is beyond one person. How Astrophage stores energy, what neutrinos have to do with it and how phototaxis works remain unresolved by the end of the book: the “molecular biologist’s” work is mostly physics, and Astrophage remains a black box. Online: yes.

S66. No backup briefing in case of amnesia. Stratt wipes Grace’s memory, and the plan rests on Yao and Ilyukhina explaining the mission. There isn’t even a video on board saying “you are Grace, here’s the task.” The ship’s procedures (airlock, centrifuge, spin drive, EVA) are stored as memories of events, not as motor skills, and the drug erases them: the very first mistake could be the last. Online: yes.

S67. Grace’s selection contradicts crew selection. A conflicted “coward,” sent by force and hating the leader, is put in a crew of three for years of autonomous work. That is a risk of sabotage and suicide, and Stratt’s bet pays off only by the author’s will. Online: yes.

S68. Nobody analyzed Rocky’s blood. Grace has a full lab and samples of Rocky’s silvery blood, but he doesn’t analyze them. “Mercury” is only Rocky’s word, and mercury blood works poorly physically. Cells in mercury float up with a force of ~12 times their own weight (~25 at 2 g), mercury dissolves aluminum, zinc, tin, gold and silver, and its volumetric heat capacity is half that of water, making it a poor heat carrier for steam-driven muscles. Online: no.

S69. Fast, precise movements on heat-driven muscles. Steam muscles are inertial, yet Rocky speaks while assembling delicate mechanics. That is possible only with a very fine vascular network, of which there is no mention. Online: no.

S70. The “powder keg” at Baikonur. 1 mg of Astrophage is ~21 t of TNT, a visible grain, easy to tell from an invisible nanogram. The warehouse issues such charges without labeling, and “no debris” would require ~100 times more energy. Online: no.

S71. The investigation is run by an interested party. The explosion removes the prime and backup scientists and frees the spot for Stratt’s candidate. She herself runs the two-day investigation, without Grace and without independent review. Online: no.

S72. Authority ends at the moment of launch. Stratt’s emergency power ends at launch, while the crisis peaks afterward: famine, farms, receiving the Beetles. No successor is named. Online: no.

S73. The common root of the gaps: power without a counterweight. All the unraised alternatives (Venus, Plan B, communication, a second ship) follow from one thing: one person makes the decisions with no scientific council and no independent expertise. Crisis projects pursue several paths at once (the Manhattan Project went through both uranium and plutonium, under the oversight of scientific committees), while Stratt has a single bet. The only counterweight is a trial after launch, when all the irreversible decisions have already been made. Online: partly.

S74. Prime and backup scientists in the same room during a dangerous experiment. The prime scientist and his backup run a dangerous experiment together and die in the same explosion, and the flight engineer didn’t know about the test. A backup exists precisely so that this doesn’t happen. Scientists of this caliber lead groups of postdocs and set tasks for the test service; they don’t run a destructive experiment in person. Their death is needed only to make room for Grace. DuBois and Shapiro are an open couple (pp. 307–309): even without an explosion, losing one would incapacitate the other, and Stratt, who feared mixed crews (p. 289), allowed such a couple. Online: yes.

S75. Medical service with no investigation of the deaths. The bodies of his comrades lie next to Grace for years, yet neither he nor the medical station establishes the cause and time of death. His decision to go back into the coma depends on it. Grace jettisons the bodies in the first hours, although tissue samples could have been taken in half an hour, and whether he survives the coma on the return trip depends on the cause of death (p. 471). Online: no.

S76. One ship and three on board — an architecture with no redundancy. Three people with no doctor and no duplicated specialties for four years of autonomous work, and one ship. Two or three ships would raise the chance of success from 30–50% to 51–88%. A second ship is never discussed in the book. Scouts and relays based on the Beetle (~0.01% of the fuel), launched years before the ship, would have checked the target and the route. Any serious mission has a ground twin of the spacecraft. An extra person is 1.5–3 t of dry mass and 30–60 t of Astrophage, two extra people are 3–6% of the fuel, and with a coma mortality of 2 out of 3, the probability that at least one survives is ~70% with three and ~91% with six. Four Beetles carry the data home: redundancy where it is cheap, and none where a failure means the end of the mission. Online: partly.

S77. Mission profile: fuel stockpiled on Earth instead of Venus. Collection at Venus would yield fuel in months, and orbital farms need 15–20 times less area. The Sahara is the worst of the options, and it is the one chosen. A tanker refueling with its catch in the breeding layer spends ~0.4% of its mass and a few days on an Earth–Venus–Earth trip, and ArcLight has already shown collection in under a year (pp. 36–37): if fuel is obtained within months, the argument against a second ship (S76) also falls away. An orbital farm at ~0.4 AU costs ~10¹¹ m² of film and ~1 million t of cargo instead of glass. Online: partly.

S78. A fueled ship in a decaying low orbit. A fueled ship holds ~2·10²³ J, the order of the Chicxulub impact. It sits in a decaying low orbit above populated land, although its own spin drives could raise the orbit in minutes. An orbit inclined at 51.6° passes over most of the inhabited land, the altitude margin is about three weeks (p. 420), and Soyuz craft dock nearby. Stratt herself says after Baikonur: “We just don’t have the procedures or experience to manage Astrophage safely.” Online: no.

S79. No Astrophage launch vehicle invented. A thermal Astrophage rocket would put payload in orbit in a single stage on ~0.4 mg of Astrophage per tonne. Instead, there are 16 conventional launches. An Astrophage thermal rocket with vₑ = 10–20 km/s reaches orbit in a single stage with a mass ratio of 1.6–2.6 against ~8.5 for chemical rockets, and the argument “it’s dangerous near Earth” doesn’t work when a fueled ship hangs in low orbit (S78). An elevator on jet-supported suspensions runs up against not energy but reaction mass (84,000–420 t per day), and the workable options, a space fountain or a Lofstrom loop, are not considered in the book. Online: no.

S80. The discoverer wasn’t brought in. The line, the project and the instrument are named after Petrova, yet she is neither in the Taskforce nor in the project. The only person with pre-dimming data wasn’t called in. Online: no.

S81. Launch from low orbit: exhaust aimed at Earth. Accelerating in a straight line from low orbit puts ~8·10¹⁹ J of exhaust (~20 Gt of TNT) onto Earth. A safe departure trajectory is not discussed in the book. Online: no.

S82. Spin drive: one design with no alternatives worked out. The spin drive is one design, one test at 1/1000 of the power lasting 100 µs. No comparison of options, no endurance tests. Online: no.

S83. Beetle tests in LEO and handing out milligrams. The “Pete” test in low orbit is a beam of ~40 kt of TNT per second above Earth. Milligrams of Astrophage (~21 t of TNT each) are handed out with no protocol. If “Pete” launched fully fueled (125 kg, about 2.7 Gt of TNT), a launch vehicle failure would have caused a catastrophe on the scale of S78, yet “a whole SpaceX launch” is presented as routine. Online: no.

S84. The breeding cycle in a blackpanel drives cells to the cold side. A cell can charge only at foil above 96 °C, but it goes to the cold glass (~50–80 °C) to divide. The foil itself glows at 4.26 µm brighter than the filter: either the filter is useless, or this contradicts S46. CO₂ heated above 100 °C under the glass itself glows in the 4.3 and 15 µm bands, and the book doesn’t explain why the cells don’t take this as a migration signal. Online: no.

S85. The Sahara farm worsens the cooling. The energy stored in fuel and leaving with the ship is removed from Earth’s heat balance: ~2 W/m², about 9% of the deficit from the dimming. The farm worsens the cooling. If the farm’s harvest is burned on Earth rather than exported, that is already heat transfer: to cover −24 W/m² over the world’s farmland takes ≈130 t/year with output of ~365 t/year at Redell’s rate, preferably in greenhouses. This option is not considered. Online: partly.

S86. The amnesia profile is tailored to the narrative. Forgetting one’s own name while keeping skills and all of one’s science is the profile of dissociative, not drug-induced, amnesia. Memories arrive as ready-made scenes, in order, and exactly at the right moment of the plot. Online: no.

S88. Venus’s phosphorus — the breeding ceiling and an exported resource. The phosphorus in Venus’s clouds suffices for Astrophage for weeks to decades, and the cells export it to the Sun. The population ceiling is set by phosphorus, not by 10% of luminosity. Fixing N₂ requires molybdenum or vanadium, which are almost absent from Venus’s atmosphere, and replenishment of phosphorus from the surface is orders of magnitude slower than the demand of 3·10¹¹–3·10¹² kg/year. Online: no.

S89. The Astrophage genome was read — and not used. The Astrophage genome was read but not used: no battery, no phage or antibiotic, no farm strains, no predator of their own. Online: no.

S90. Grace’s diagnosis (GCR) explains events worse than head-on gas. Galactic cosmic rays are isotropic, and you can’t hide from them behind tanks. Over 46 years Rocky would have accumulated 10–20 times more than the crew. The crew’s deaths are better explained by head-on gas at ≳0.4 c. If the crew was killed by head-on gas (~1 Gy/s from 0.4–0.5 c), a human without an Astrophage layer would have died the same way, and the moral “they didn’t know radiation, but we do” is false. A hull of xenon and Rocky’s mercury blood are the best targets for spallation (~20 neutrons per proton in mercury against 1–2 in water), and mercury also becomes activated (Hg-203, 46.6 days), so Rocky’s survival is even harder to explain. Online: no.

S91. Radiation resistance of xenonite. Xenonite, a xenon compound made from “liquid and liquid,” would pick up ~10⁷ Gy over the flight and lose 10–50% of its bonds. Its radiation resistance is unexplained, and the Eridians couldn’t have tested it. Xenon compounds are chemically weak (XeO₃ is an explosive), and radiolysis at ~3·10⁶–3·10⁷ Gy would release gaseous xenon and an oxidizer inside the hull. It is ionization that does the damage; nuclear displacements are small (~10⁻⁵–2·10⁻⁴ per atom). Online: partly.

S92. Rocky and the ammonia in Grace’s compartment (pp. 354–362). Ammonia from the airlock would reach 15–40% in Grace’s compartment against a lethal 0.5–1%. Rocky’s body at 1 atm would boil, and equalizing pressures through an opening is a shock wave. Nobody would survive. The shock wave at 29 versus 1 atm is M≈2.4, ≈5.7 bar behind the front and up to ≈27 bar reflected, while 3–5 bar is lethal for half of humans, and venting the airlock into space is explosive decompression for Rocky without his ball (water at 210 °C flashes to steam by ~35%). A variant where both survive (Rocky in the ball, the ammonia pumped out in advance) is not described in the book. Online: no.

S93. Astrophage is an automaton with memory; nobody checked the state of the fuel. A cell’s behavior is set by its history (after division, toward the bright; hungry, to charge), that is, it is an automaton with memory. The state of the fuel in the tanks over years of storage was never checked. An evolved automaton almost certainly has a migration timeout, and over years of storage the fuel would have switched to “charging” and stopped responding to 4.26 µm; no one measured the thresholds and timers. Online: no.

S94. “Exhaust all energy” vs. a cell with memory. A cell with a migration program keeps a reserve for the next stages and doesn’t burn down to zero. And the “waiting for launch” stage is a ready-made safe storage mode that the book doesn’t have. A cell’s reserve with memory is ~1–2% of its energy with a return by falling, ~6% with the canonical return flight and ~28% with braking at the Sun; the fuel calculation is overstated by that fraction. The opposite is also impossible: fuel that gives up no more than half its mass yields a ceiling of ≈0.33 c with braking. Online: no.

S95. No tests of the strain, and the Venus analog is incomplete. The Venus analog has no acid aerosol, no UV and no real cold. There are no full tests of the strain, so the side effect, passing through xenonite, is found by accident. The 0.02 atm layer on Venus is ≈72 km, ~−45 °C, a haze of 75–85% sulfuric acid, and the eating rate at real Astrophage density and 12 years in the Beetles’ mini-farms were not tested. There are sequencers on board (p. 153), but no one reads Taumoeba’s genome. Online: no.

S96. The Threeworld margin is 3% (pp. 453, 459). Taumoeba’s limit is 8.25% nitrogen, and on Threeworld it is 8%. A margin of 3%, though Rocky himself proposed Taumoeba-86. Online: no.

S97. Fuel with no safety catches. The fuel is a paste at 96 °C with CO₂ and is always armed. A nuclear charge has several independent interlocks, but here is a jar with a photon bomb and not a single safety. One sated cell (≈1.5 MJ) heats ~4 kg of paste by 90 K, and an opaque partition would simply heat up and glow onward, so the only realistic options are stage blocking and cold (glow at 4.26 µm drops ~40 times at −50 °C), both active. Taumoeba is unsuitable as a “fuel killer”: it eats for hours, while cells discharge in fractions of a second. Online: no.

S98. Rocky “can’t” make a sampler for 46 years, then makes everything in days. For 46 years Rocky “can’t” make a sampler, and then in days builds a sphere, ten tanks with 1 ppm dosing and copies of the fuel tanks. A superalloy “harder than any metal alloys” (p. 474) can’t be obtained from an unrefined “metal stew” of the ship’s melted-down systems (p. 457). Online: partly.

S99. The Hail Mary wasn’t reinforced with xenonite (pp. 448–449, 451, 457, 464). Xenonite is available, yet it was used for neither the tanks, nor the head-on gas shield, nor the heat shield, nor the partitions between compartments. The selection tanks are made of xenonite, and a control slide behind a xenonite membrane in each round of selection would have caught the passage at once. Online: partly.

S100. The Blip-A holds the planet’s full pressure. The Blip-A hull holds 29 atm, though ~20 atm would suffice for the body. If H44 is right, that is a 2.5-fold overspend of hull mass. Online: no.

S101. The medical bed’s secret ground-loaded program. The medical bed hides a procedure from Stratt’s people, a DGSE dose that bypasses the commander. A channel of interference with the coma automation, the only candidate for the cause of the crew’s deaths, is established by the text. Online: no.

S102. Stratt’s plan depends on a living crew. There is a safeguard against 1% disloyalty from Grace, but for the far likelier death of the crew in the coma (1 of 3 survived) there is neither a recording nor an instruction. Online: yes.

S103. Grace’s qualification is self-fulfilling. Grace is taken for his thesis “life without water,” and twenty pages later he himself proves that Astrophage is mostly water. The basis for the recruitment vanished, and Stratt spent years creating the qualification. The “life without water” thesis is refuted twice (Astrophage is “mostly water,” p. 71; Erid has liquid water, p. 328), yet at Erid Grace rejoices: “it proves I was right!” (p. 171). Years of teaching mean no publications and no fresh methods, and Grace’s “biology” on board is microscopy and stepwise selection, the level of a lab class; he does no genomics, biochemistry or physiology of Astrophage. Online: yes.

S104. Taumoeba’s path and motive on the Blip-A. Selection was for “burrowing into the wall,” not for passing straight through. Outside Rocky’s farms it is 210 °C, and Taumoeba dies above 107 °C. “Nothing standing between” is just Grace’s guess. Online: yes.

S105. The sacrifice isn’t necessary: Rocky only needs fuel for braking. The Blip-A is intact and flying to Erid, and Rocky lacks only ~50–70 t for braking, ~3% of Grace’s fuel. Casting a steel tank and topping up takes days; no sacrifice is needed. Online: no.

S106. Grace doesn’t look for food himself. A biologist spends six weeks planning death by starvation and doesn’t think of Taumoeba as food; the solution comes from Rocky’s remark. Nobody checks the monodiet for vitamins, amino acids and toxicity. Online: no.

S107. Solar photometry by a blind civilization without electronics. To say “full luminance,” you need photometry better than 0.5% over decades. A blind civilization without electronics can’t do that. The absolute route requires the Sun’s parallax (≈40 mas from Erid’s orbit) with an accuracy of ~0.1 mas, the relative route requires instrument stability of ~0.5% over decades, and without electronics only thermocouples with a galvanometer remain, as for astronomers of the 1920s. Online: no.

S108. 16 years at 2.1 g for a man over 50 — underestimated (refines minor item 1). At 2.1 g the pressure at the brain is ~47 mmHg (fainting), and a fall from standing height is like falling from 3.7 m. For an emaciated 50-year-old, 16 years of this is a serious threat the book doesn’t notice. The pressure in the veins of the feet is ≈190 mm Hg (edema, varicose veins), and scurvy weakens the collagen of bones and blood vessels, so the risks add up. The obvious ways out, living on the Hail Mary with a 1 g centrifuge above the radiation belts, a pool, a lying-down regime, are not considered. Online: partly.

S109. Eridian food could have been purified. Proteins and sugars are shared by humans and Eridians, and heavy metals separate with school chemistry: sulfides, chelates, ion exchange. That would have removed both the “sacrifice” and the hunger on Erid. Online: partly.

S110. A cell’s flight between stars is not justified energetically. The cell is a photon rocket: accelerating to 0.9 c and braking need 18–23 times its dry mass in fuel and 4–5.5 years of charging at 6 mW. It has to brake blind ≈2.4 light-years from the target, when ≈10⁻¹⁷ W arrives from the star and the sky is distorted by Doppler (×4.9) and aberration (half the sky in a cone of ≈23°). The star’s gravity won’t catch a cell at 0.9 c. The agent called this the book’s main stretch at the time. S55 is about the flight not paying off; this item is about whether it is physically possible. Online: partly.

S111. There is no evolutionary path to the Astrophage cycle. “Half” of a light thruster and “half” of a neutrino battery are useless, so selection has nothing to build on. The book presents the cycle as a given. Online: partly.

S112. Nobody notices the oxygen on Venus. The cells build themselves from CO₂, and the oxygen stays in the atmosphere: ≈5·10¹⁴ kg of O₂ per year against a current level below 1 ppm. Spectroscopy would see this within years, yet in the book Venus is studied and even sampled from orbit. Online: no.

S113. Astrophage DNA near a star without repair. Gamma quanta and cosmic rays with energies on the order of GeV pass right through the cell, and DNA at 96 °C near the photosphere survives only on the assumption of perfect thermoregulation. The book has no mechanism of protection or repair. H22 only covers the contrast “a needle kills, a proton doesn’t.” Online: partly.

S114. A hibernation gene doesn’t help in a medically induced coma. By the book, coma resistance is a legacy of ancestral hibernation. But in hibernation metabolism drops tens of times, while in a medically induced coma it is almost normal, so hypometabolism genes give nothing over 4 years of coma. Besides, the 1/7000 frequency would differ between peoples, while in the book it is the same for the whole world. Online: partly.

S115. Reinfection from neighboring stars after the “victory”. By the book, neighboring stars are infected and spores fly from them to the Sun. Taumoeba at Venus dies out without food, while the infection will keep coming back, and the epilogue provides for no permanent reseeding of Taumoeba. The Sun’s final recovery is not justified. Online: partly.

S116. Rescue arrives after the peak of the disaster. The Beetles’ answer arrives no sooner than ~26 years after launch, plus time to breed Taumoeba and deliver it. The Sun loses 10% in about 25 years, and Leclerc predicts the death of half of humanity in 19. At best the mission treats the consequences, and with such timelines a plan B is mandatory (to S40, S54). Online: partly.

S117. The IRAS, WISE and JWST surveys didn’t find the Petrova line earlier — and it should vanish for months. The sky at ~26 µm was observed by IRAS (25 µm), WISE (22 µm) and JWST/MIRI. They don’t observe near the Sun, but the line’s light scattered by dust at large angles should have been noticed, which conflicts with the brightness estimate (S6). Besides, the arc turns with Venus, and for several months a year, while Venus is near the Sun, the line cannot be seen from Earth. The book has no such seasonality. Online: no.

S118. A continuous arc although the cells glow only when accelerating and braking. If the cells glow only when they fire their engine, the arc between the acceleration and braking zones should be dark. A continuous line requires either glowing while coasting or different cells switching on in different places. The book says neither. The same contradiction appears in the scene at Adrian, “the whole screen is red, we’re inside the line.” Online: no.

S119. The word “neutron” never appears in the book. Besides Venus (H48) and the bombs (S26), neutrons are missing in three more places. Head-on protons of 1–1.5 GeV knock secondary neutrons out of the hull, and these upset the electronics (to H12). The Hail Mary’s nose becomes activated to ~10–50 mSv/h (Na-22, Na-24), and the Geiger counter on board doesn’t check the hull. Eridian metallurgists, with atomic masses ≈2Z and fractional masses (chlorine 35.45), were bound to arrive at isotopes (to S18). Online: partly.

S120. The mechanism for turning the cell is not described, and rigid thrust is unstable. The book only has “stimulus — response” behavior. If the thrust is rigidly tied to the body (and the spin drive requires that), at 10⁴ m/s² an offset of 1% of the radius spins the cell up in fractions of a millisecond. Stabilization is needed within microseconds, and biochemistry can’t do that. A consistent model, the cell as a spherical microlens at 4.26 µm emitting from the focal spot, always gives thrust toward the source, and such a cell at Venus can only accelerate toward it, not brake; at 18.31 µm it is too small to be a lens. Online: no.

S121. The Taumoeba farm in a Beetle sits next to its own fuel. The Beetles fly on Astrophage, which Taumoeba eats, and strain 82.5 has already shown that it passes through xenonite. The farm’s isolation relies on an aluminum hull — the same weakness that ruined the Blip-A’s fuel, and over 12 years of flight nobody will check it. Online: partly.

S122. The Eridians have a crew of 23 without hibernation, Earth has three in a coma. “Object A” carried 23 awake Eridians with food for the round trip and a stock of xenonite. On the same fuel, Earth builds a ship for three in a medically induced coma, one way. The book doesn’t explain the difference. Online: partly.

S123. A chain instead of a cable or rods, although the Eridians have xenonite elevator cable. Rocky’s civilization has long made space elevator cable 17,300 km long, yet at Adrian Rocky and Grace spend two weeks casting 200,000 links by hand. Any driller would join rigid rods: fewer parts, no sagging in the reflected exhaust, and no links scattering into the beam (H46, H54). The excuse “I can’t build a reel for the whole chain” doesn’t work: a cable winds tightly. The chain is needed for a scene of working together. Long rods with eyelets and universal-joint pins could be cast: the suspension is flexible, works in tension rather than bending, as the oval links do, and has several times fewer parts. Online: partly.

S124. A separate sampler was not considered. Instead of holding an interstellar ship on thrust above the atmosphere, a drone with an aerodynamic skip or a small return capsule with a heat shield could have taken the sample. Then they would have risked an instrument, not the ship. The Hail Mary has no standard means of collecting samples at all. Online: partly (no standard probe).

S125. Rocky is not an engineer but a talking CNC machine. In the plot Rocky makes anything to order within days (tanks, a sphere, a chain, mini-farms), but he designs and optimizes nothing himself. An engineer would have chosen a cable or rods, not a chain (S123). He carries out other people’s ideas, which fits poorly with the role of the expedition’s engineer (to S98, X1). Rocky builds drones as needed, and the casting cycle (dosing, curing, removal, joining) is simpler than autonomous flight, and with 200 thousand repetitions a mechanism pays for itself; the book doesn’t say why this can’t be mechanized. Online: partly.

S126. The Beetles weren’t given the protection the ship was given. The Hail Mary’s crew is shielded from the head-on gas by a 1 mm layer of Astrophage (Lokken’s idea), while the Beetles fly unprotected. So the problem was understood and deliberately left unsolved: the probes are known to be inoperable (H12). This is one of the pieces of evidence for X2. Online: no.

S127. Rocky bypasses a Beetle’s computer controls in an hour. “Eridian electronics isn’t nearly so advanced” — and right away “An hour later, he’s bypassed all the computer controls” (p. 412). An engineer without transistors or computers masters an alien digital system in an hour. Complements H31 and H32 and serves as evidence for X1. Online: partly.

S128. A vacuum gap instead of xenonite. Taumoeba can’t move through vacuum, so a double wall with a vacuum gap is simple and reliable protection for the fuel. Instead they install xenonite, which the strain passes through. Online: partly.

S129. The dimming of neighboring stars should differ by orders of magnitude, and a star without a CO₂ planet doesn’t dim at all. With exponential growth, a difference in infection time of a few years or decades gives a difference in dimming of orders of magnitude, yet the book has no spread. A star without a CO₂ planet gives Astrophage nowhere to breed: it doesn’t dim and cannot be a link in the infection chain. And if the ceiling is set by the composition of the incubator planet (phosphorus, trace elements, S88), it should itself differ from star to star, rather than being the same 10% (see S17). Online: partly.

S130. Diving into the Sun: the limit is the cell’s capacity, not temperature. The book assumes Astrophage tolerates any heating and sets no limit on its energy reserve. In the photosphere (5800 K) a cell receives about 0.02 W and collects a mass equal to its own in about 26 days; at a depth where it is ~10⁵ K the inflow is about 180 W, and the same reserve is collected in minutes. A cell that dives a little deeper would overflow almost instantly, and what happens on overflow and what pressure it withstands the book doesn’t say. Online: no.

S131. A precise 96.415 °C thermostat is impossible. According to the book, neutrinos are produced in collisions of thermal protons above an exact threshold, and the cell holds exactly 96.415 °C. But Maxwellian speeds are distributed smoothly, fast protons exist at any temperature, and neutrinos would be produced below the threshold too, only more rarely. There is no sharp threshold and no precise thermostat (see H14). Online: partly.

S132. In the lab the cells didn’t divide for a week, although both division stimuli were present. According to the book, division is triggered by CO₂ light plus contact with CO₂. The lab has both CO₂ (~400 ppm in the air) and IR glow of the walls at 4.26 µm, yet the cells didn’t divide for a week and divided only after a dash to the filter. ArcLight could only have brought cells “bound for Venus” before division, and if they were taken already in the breeding layer, at the target, they have no reason to “look for Venus”: the stimulus model doesn’t add up (see S46). Online: partly.

S133. Astrophage makes humanity multiplanetary, and the book doesn’t see it. By canon, Astrophage simultaneously gives powerful interplanetary transport, energy for settlements and radiation shielding, that is, it removes the main barriers to settling the Solar System. Terraforming Venus with it is hard (atmosphere ~4.8·10²⁰ kg, ~92 bar, 464 °C, leftover oxygen, little hydrogen), but as a fuel source and an industrial base with stations it is realistic. This noticeably weakens the “catastrophe with no alternative” scenario (see M1). Online: partly.

S134. Greenhouse measures cover only a small part of the dimming. Raising CO₂ from 420 to 1500 ppm gives ΔF = 5.35·ln(1500/420) ≈ 6.8 W/m², while losing 10% of the light is ~24 W/m²: CO₂ compensates only about three percentage points. Industrial HFCs, PFCs and SF₆ don’t scale linearly with GWP (the bands overlap and saturate), and long-lived gases will create a new problem after the Sun recovers. Recirculating heat through Astrophage adds no energy: real heating needs Astrophage charged outside Earth (see M2). Online: partly.

S135. SF₆: “thousands of times” is an underestimate, and the main argument is missed. To give 12–24 W/m² you need 0.5–1 Gt of SF₆, with world output of ~10 kt/year: 5·10⁴–10⁵ times more, not “thousands of times.” The real argument against SF₆ is its lifetime of about 3200 years, that is, irreversibility, but it isn’t in the book, even though reversibility is exactly what Leclerc cites in favor of methane (see S26). Online: no.

S136. There is no way to harvest from sealed panels. A panel is closed with glass and yields about 0.18 mg of Astrophage per year. Opening each one once a year means ~5.5 billion panels per day and ~11 million workers on harvesting alone, even at one minute per panel, without dust and repairs. The alternative is drainage or pipelines to each panel (~0.5 kg per day per km²): a network of millions of kilometers plus CO₂ and water supply, and none of this is in the book (see S15). Online: yes.

S137. There is no way to supply an army on the Sahara farms. The entire Sahara has 2–3 million inhabitants, while the farm needs 10–35 million (with robots) or 100–300 million (by hand). At 80 L of water per person per day that is 1–12 million m³ per day, 1–12% of all the world’s desalination plants, with water mains of 1000–2000 km; food is 30–300 thousand t per day (1–12 thousand trucks a day) with almost no railways, about 900 settlements, and in summer at 45–50 °C work is possible only at night. The infrastructure alone will take years, more likely a decade, and doesn’t fit into the book’s “a few years” (see S15, S41). Online: partly.

S138. The largest project is run by a convicted embezzler with no audit. The Sahara farm is run single-handedly by Redell, a convicted embezzler with a gambling addiction, and audits and oversight are not mentioned in the book. His accident (the real system launched as a “virtual test,” no interlock, seven people killed) is later repeated in the deaths of DuBois and Shapiro (S70, S74) and built into the “Pete” test (S83). This is a failure of management within the logic of the world, not a genre complaint (see S73). Online: no.

S139. Stratt deceives the commander and rejects a volunteer. Yao states plainly: “I will not have someone on my crew who is there against their will” (p. 422), and Stratt deceives him with the story that Grace himself asked for the anesthesia (p. 429). The volunteer, Paraguayan chemist Andrea Caceres, known “back during the first call” (p. 421), is not trained, even though among tens of thousands of volunteers, with a gene frequency of 1 in 7000, there were several carriers; and within two days of “investigation” Stratt already has a checklist in Dutch, a soldier at the door, a DGSE drug and a program for the medical station. Stratt’s remorse is known only from her own words, and humanity’s key mission is entrusted to a person who deceives her own crew commander (see S37, S71, S103). Online: yes.

S140. A suicide crew: means of death instead of psychological support. Stratt cites psychiatrists: “crushing depression is likely. And suicide is a real risk” (p. 122); DuBois stockpiles nitrogen and writes a note, Ilyukhina takes heroin (p. 374), Yao asks for a pistol (p. 375), Grace goes through the methods (p. 113). Selecting those “willing to go on a suicide mission” from a pool where the “insane or suicidal” are, according to Stratt, “probably a lot” (pp. 278–279) skews the sample toward people with depression and a taste for risk. All psychological support is reduced to means of death (see S67). Online: partly.

S141. 2 g of Astrophage on an aircraft carrier: three Hiroshimas with X-rays. Dimitri is given 2 g of Astrophage for the test: 1.8·10¹⁴ J, about 43 kt of TNT. “300 km, so is okay” saves the coast from a tsunami, but not the aircraft carrier with ~5000 people: 26 µm light is absorbed in the first meters of humid air, ~3·10¹⁰ J/kg means plasma, soft X-rays and UV, then a shock wave, as in an air nuclear burst without neutrons. Moreover, by the book’s own logic (S46) the vial at 20 °C itself glows at 4.26 and 18.31 µm, so such a container was not safe (see S70, S83). Online: no.

S142. Two-body photon recycling is not considered. Inside a single ship, absorbing your own exhaust is pointless: the absorber receives the reverse momentum, and the thrust drops to zero. But between a ship and an “anvil” (a station, a jettisoned mass), Astrophage on both sides could bounce the same light back and forth, multiplying the thrust by the number of reflections (a real idea, the photonic laser thruster, Bae). For accelerating from an orbital station or at Venus this is a strong scheme, and it isn’t in the book (see S79, S77). Online: partly.

S143. Gene transfer from the farms: a biothreat no one sees. With 98% similarity to Earth life, genes from trillions of cells on the Sahara farms could pass to soil bacteria, and an organism that absorbs heat and glows in the IR would become a new pest. Over the years of the project the book gives not a single genomics result, although in the 2020s a single cell’s genome is read in days, and on board there are sequencers and a backup specialist in DNA splicing (see S89). Online: partly.

S144. Dust on the route: a 1 µg dust grain at 0.9 c is about 30 kg of TNT. The book protects the ship from head-on gas with a 1 mm layer of Astrophage, but doesn’t discuss dust and small clumps. A 1 µg grain at 0.9 c carries ~1.2·10⁸ J, and even a 1 µm grain about 3 kJ; such clouds aren’t visible from Earth, only a scout would reveal them, and there is none. A cloud can’t be avoided at 0.9 c; one can only slow down in advance (see S7). Online: yes.

S145. A light report from Tau Ceti would have arrived a year before the Beetles. While the Hail Mary accelerates, its exhaust points at the Sun, and modulating the thrust would give Earth “ship alive” telemetry for years (the receiver retunes for Doppler: ≈45 µm at 0.5 c, ≈113 µm at 0.9 c). From Tau Ceti one spin drive element out of 1009 would suffice: ≈9 TW, ≈8.6 kg of Astrophage per day, tens of bits per second; light takes 11.9 years to reach Earth, while the Beetles fly ≈12.8 years, so a report with the Taumoeba selection recipe would have arrived a year earlier and insured against probe failure. The transmitter (the spin drive) and the receiver (the Petrovascope, p. 116) are already on board, yet no one discusses light communication while planning the mission (see S60). Online: yes.

S146. “Why now”: if Astrophage is ancient, Tau Ceti’s neighbors would always have been dimming. Grace himself allows that an ancestor of Astrophage seeded Earth and Erid with life billions of years ago, so spores have been flying between stars for a very long time. A star reached by Astrophage loses 10% within decades, so neighbors should have been dimming constantly, over geological eras, not in a wave in recent decades. An explanation (a new mutation giving a range of 8 light-years, or the disappearance of predators at neighboring stars) is not in the book (see S39). Online: yes.

S147. Tau Ceti’s long-standing plateau could have been checked from Earth. Systematic photometry begins with photographic plates of the 1890s with an accuracy of ~0.1 magnitude, i.e. about 10%: a star that dimmed earlier “has always been like that” for us. But Tau Ceti’s radius has been measured by interferometry with an accuracy of ~0.5%, from radius and temperature the luminosity is predicted with an error of ~4%, and a 10% deficit is visible at the 2–3σ level. Given the “Tau Ceti is the homeland” version, the “long on the plateau” variant is a priori more likely than the book’s, and the resolution rests on luck that the heroes don’t recognize as a risk (see S39a, S40). Online: partly.

S148. Rocky notices the Hail Mary only with lucky geometry. The Hail Mary’s braking exhaust points at Tau Ceti, and from the side, in vacuum, a narrow beam isn’t visible; Grace himself makes the reservation: “depending on where they were.” The chance of happening to be in the beam cone is ~2·10⁻⁵ for a 1° width and ~2·10⁻³ for 10°, and the ship’s light scattered on dust (~5·10¹⁴ W against ~10²¹ W of flux toward Venus) is ~10 million times weaker than the Petrova line. The first detection rests on luck (see S30). Online: partly.

S149. Rocky hears Grace through two floors. “Rocky is two floors below me, but I speak at a normal volume. I know he can hear it just fine” (p. 302). Through one tunnel wall speech would be attenuated by about 30 dB and still be intelligible, but through floors two levels down, not. Together with the light-guide nerves (p. 265), the “old technology” camera and the “didn’t notice the light went out” scene (p. 396), this points to a hidden perception system rather than passive sonar (see S33, X1). Online: partly.

S150. A chemical rocket can’t launch the elevator seed from a super-Earth. An elevator is built from the top, from synchronous orbit, and even under the Edwards scheme a seed ribbon of ~20 t launched by rocket is needed. To Erid’s synchronous orbit Δv ≈19 km/s without losses; on hydrogen and oxygen (Isp 450 s) the mass ratio is ≈74, with four stages the payload is ~1/200, i.e. a rocket of ~4000 t (≈1.3 Saturn Vs), at Isp 350 s ~25,000 t, and at 29 atm the specific impulse is even lower (Hippke 2018). A nuclear rocket contradicts the ignorance of the nucleus (S18), an Astrophage one is a vicious circle (H19), and with such a program the Eridians would already have flown beyond orbit: from synchronous orbit to escape velocity only ≈4.4 km/s is missing (see S19). Online: partly.

S151. Erid’s elevator cable requires xenonite at the level of ideal nanotubes. The ratio of cable thickness at synchronous orbit to that at the surface on Erid is ≈9 for ideal nanotubes (46 MJ/kg) (2.9 on Earth), and ≈6·10¹¹ for Zylon (3.7 MJ/kg). So xenonite’s specific strength is tens of MJ/kg, the level of theoretical nanotubes, and with such perfection over 17,000 km of a material that is cast from two liquids, like epoxy. A super-Earth is the worst case for a space elevator (see S5, S19). Online: partly.

S152. Loads on Erid’s elevator cable: wind, Coriolis, stellar tides. The dynamic pressure of a 10 m/s wind in ammonia at 29 atm and 210 °C (≈12 kg/m³) is ≈620 Pa against ≈60 Pa on Earth, and the main load is carried by the lower 50–100 km of the cable; add lightning and corrosion in hot ammonia. A day of 5 h 6 min gives strong Coriolis effects and deflection of the cable by the climbers, the star at 0.2 AU gives tides, and the cable has its own oscillation modes. How the anchor, oscillation damping and wind protection work, the book doesn’t say (see S19). Online: no.

S153. Erid’s oblateness is not taken into account. With 8 Earth masses, 2 Earth diameters and a day of 5 h 6 min, the ratio of centrifugal acceleration to gravity at the equator is q ≈ 0.076 (Jupiter’s is 0.089), hence an oblateness of ≈0.07–0.09: the equatorial radius exceeds the polar one by ≈900–1200 km, gravity at the equator is 10–15% lower than at the poles, and “2 g” is either an average or a value at one point. A short day gives narrow latitudinal jet streams, as on Jupiter, and with an atmosphere 29 times denser than Earth’s this is a serious lateral load on the lower section of the elevator cable. The book is silent about oblateness (see S20). Online: no.

S154. Erid’s elevator climber: energy and power electronics without transistors. Lifting to Erid’s synchronous orbit (≈17,300 km) costs ≈100 MJ per kilogram of cargo (≈48 MJ/kg on Earth) and takes ≈3.6 days at 200 km/h or ≈17 hours at 1000 km/h. Energy must be supplied through the cable or by beam, and power electronics on that scale for a species without transistors is a separate unsolved problem the book is silent about (see S19). Online: no.

S155. Erid’s elevator passes through the hottest part of the radiation belt. If Earth’s magnetosphere is scaled to Erid’s field (≥25 times Earth’s), the belts will occupy ~1.2–12 planetary radii, and synchronous orbit at 2.36 radii lies in the inner belt. The elevator runs along the magnetic equator, where the particle flux is greatest, the belts can’t be bypassed via the poles, and X-rays and flare protons at 0.2 AU are about 25 times stronger than at Earth. Eridians working for years in orbit and on the elevator could not have avoided encountering radiation (see S18, S21). Online: yes.

S156. Collective “humming in a trance” as a computer. Eridian scientists solve complex problems by collective “humming in a trance,” and they calculated the route and the flight time (6.64 years) without transistors or computers. The arithmetic could still be explained by an exceptional memory, but a collective trance as a computing machine has no justification at all; in essence it is magic. Yet since there is radio, there are electromagnets, and relays are made from them: the Z3 relay computer (1941) is technology of the same level (see S21, H31). Online: partly.

S157. The exchange of knowledge is one-way. About his main technology, Rocky the engineer says “I not understand science. I just use” (p. 285), the whole xenonite recipe is “liquid and liquid. Mix” (p. 325), and Grace drops the subject. Grace gave away special relativity, the physics of radiation, how computers work, Wikipedia, Earth’s location, human biology, Taumoeba and, in the epilogue, himself, while Rocky gave only products and labor, not a single theory. Meanwhile his account of breeding Astrophage in the ocean contradicts H17, H27, H28 and H29 at once (see X1, S127). Online: partly.

S158. Rocky has about 2% living tissue, and Astrophage already serves as his coolant. According to the book (p. 265), Rocky has a carapace of oxides, bones of alloys, mercury blood, silicate nerves, inorganic muscles, and only “a few kilograms” of living tissue out of 168 kg. Rocky’s ball-shaped spacesuit is cooled by Astrophage (pp. 269–270), and Astrophage would also work as a battery: 300 W over 46 years is about 4·10¹¹ J, only ≈5 mg. If the Astrophage inside Rocky is a domesticated organ, H28, H29, H33 and Rocky’s survival next to the tanks are all resolved, but the book doesn’t say so (see X1). Online: no.

S159. The Eridians have enough fuel to escort an alien ship. Rocky gives Grace “over 2 million kilograms” (p. 481), the Hail Mary’s full launch load, and keeps “millions of kilograms” for himself (p. 493). At ~0.88 g, a Tau Ceti → Earth trip requires ≈160 dry masses of fuel, Earth → Erid ≈280, and the Newtonian “there and back” reserve after arrival leaves a remainder of ≈1400. Such a surplus is natural for a post tasked with escorting the alien home, and the ending carries this out literally: on the donated fuel Grace flies not to Earth but to Erid (see X1, M7, H30). Online: partly.

S160. Ammonia burns: 6 hours later Grace is already working. Grace is burned by a jet of ammonia at 29 atm and 210 °C, he is unconscious for 6 hours, and then he works (pp. 360–362). Toxic pulmonary edema from ammonia usually develops after 12–72 hours, i.e. the most dangerous part would still be ahead of Grace, yet the book has no consequences. Rocky, with three severe injuries by Grace’s own list (p. 362), also recovers without a trace (see S92, X1). Online: partly.

S161. Grace searches for Rocky over the whole sphere, although he knows his course. Grace knows Rocky’s course “down to the milli-arcsecond”; only the moment of engine cutoff is unknown, so the error lies along the track, ≤1.7 km/s × ~2 months ≈10 million km, and across it only from hundreds to hundreds of thousands of km. The right approach was to calculate the track, set out on an intercept with lead, and scan a narrow cone along it, where the echo time immediately gives the range. Instead Grace flies to a “best guess” point (p. 496) and scans the whole sphere (p. 499): an easy task turned into a hard one and solved by luck (see H63). Online: no.

S162. 16 years on Erid, not a single attempt to contact Earth. The Eridians have the Hail Mary with its Petrovascope, millions of kilograms of fuel, Wikipedia and Grace, who knows how the Beetles are built: a probe of that class would reach Earth in ~17.5 years, a beacon in the Petrova line in 16.3 years. There are motives (a second species fighting Astrophage, the threat of reinfection from neighboring stars, scientific exchange), but the largest “thrum” is convened for Grace’s life support and none for communication, and Grace doesn’t even ask. Either this is a stretch or, in the logic of X1, communication goes on without him (see S60). Online: yes.

S163. The Sun cannot recover in less than a year. In the epilogue the Sun has “fully recovered its luminosity” in less than a year. But Taumoeba eats Astrophage only in the breeding layer at Venus, while cells already at the Sun keep absorbing light for their entire charging period, which is much longer than a year (H70). Stopping reproduction doesn’t remove the existing population: an additional mechanism is needed (rapid cell death, a halt to absorption, a mass departure to Venus), and the book has none (see H26, S27). Online: partly.

S164. The epilogue describes an enclosure almost word for word. Grace knows about the Sun’s recovery only from the “Astronomy hive” (pp. 514–515), about the deaths of the Blip-A crew and the “46 years” (p. 283) only from Rocky’s words, and it is Rocky who initiates contact, immediately repeating Grace’s speed of 8.6 cm/s (p. 141). The epilogue (pp. 512–518): “I’m of extreme scientific interest,” a dome of transparent xenonite “because they thought that’s what I would want,” 30 keepers “or so I’m told,” light on a daily cycle, meat grown from Grace’s own cells, a daily showing to selected children. In this reading the motive is use, not capture: Taumoeba on Earth destroys Earth’s Astrophage energy sector too, and with it the interstellar fleet (see X1). Online: partly.

S165. From Earth the mission looks like three decades of silence right at the peak of the disaster. The acceleration is visible for years in the Petrova line (exhaust toward Earth), braking at Tau Ceti isn’t visible, then silence until the Beetles, ~26 years after launch. If the aim of sabotage is to freeze other solutions for the duration of the mission, what is needed is a ship that really flies and a crew certain to fail, and the survivor’s amnesia gives a failure without a witness or traces. Evidence: without Rocky, Grace would have died several times (S65, S66, H43, H45), and the Beetles weren’t given the protection the ship was given (S126) (see X2, S116). Online: partly.

Systemic Gaps

M1. Astrophage changes the very framing of the crisis. It takes away sunlight, but gives a way to collect, store, transport and release enormous energy in a directed way. The scenario “dimming → famine → the only hope is an interstellar expedition” doesn’t work without a comparison with local solutions (extraction, heating, greenhouses with artificial light). Online: yes.

M2. The 10% limit is a task for adaptation, not the end of the world. If the limit also applies to the Sun, the energy deficit is finite, and Earth can be warmed: raise CO₂ to ~1500 ppm, take methane from marine gas hydrates, produce freons with Astrophage energy. Methane from Antarctic ice (S26) is the worst of the options. Online: partly.

M3. Neutrino dumping as an expendable refrigerator. By canon, when a charged cell dies its energy goes into neutrinos (the IceCube scene). Then starving cells absorb heat and charged ones can be destroyed — a ready-made heat sink the book doesn’t use. Online: no.

M4. Starving Astrophage as engine protection. Starving cells could take up the waste heat of the spin drive and the optics. The book doesn’t consider thermal protection of the engine at all. Online: no.

M5. Hot environments make the solar trip unnecessary. If Astrophage charges from any heat above 96 °C, it can be grown at geothermal sources, and the natural population would settle in the hot layers of Venus and cool it. The book doesn’t explain why the niche is unoccupied (to H27). Online: yes.

M6. Managing the flows instead of waiting passively. Phototaxis makes it possible to gather Astrophage with an IR lure and lead the flow to facilities or to another planet. The lure would need 0.2–1 TW, about 1 kg of Astrophage per day (to H42). Online: yes.

M7. The Eridians’ production base is not shown. A crew of 23, a flight without hibernation, millions of kilograms of fuel: where, from what energy flow and in what time did the Eridians produce them? The words “the engineers built in a margin” are no substitute for a calculation. Online: partly.

Hypotheses

X1. Hypothesis: the Eridians lie/manipulate (dark forest, domestication). Everything known about the Eridians is known from Rocky’s words. If the Eridians are technically stronger than they show and Rocky is a sentry post at Tau Ceti, then radio without transistors, 46 years of waiting, “empty space” and the epilogue-enclosure all close at once. Online: no.

X2. The “mission sabotage” hypothesis. The same channel that gave Grace the amnesia drug could have given the crew a lethal dose. The cause of death isn’t established, there are no witnesses, and the survivor remembers nothing. There is no direct evidence, and no contradictions either. Online: no.

Genre Notes

G1. Genre conventions (folksy-print tropes, Dumas, coercion approved, deus ex machina, Robinson and Friday). Folksy-print characters, villains cast as heroes “like Dumas,” approved coercion, deus ex machina, the “Robinson teaches Friday” lesson. These are complaints about the genre and the text, not about the logic of the world. Online: yes.

G2. Two clichés: the “magic material” and the “amnesiac who lands in another world”. Xenonite closes any engineering problem, and Weir himself admits he didn’t work out its composition. The amnesia is needed to reveal the past through flashbacks and so that nobody on Earth asks an awkward question about the point of the mission. Online: yes (Kowal on the amnesia cliché, xenonite as “magic stuff”).

G3. Scientism as a costume. Harrison, Asimov, Hamilton and Lem don’t explain their assumptions, and nobody holds it against them. Weir buries himself with the words “ATP,” “mitochondria,” “neutrinos” — like Penny from “The Big Bang Theory,” who puts on glasses and languidly says “molecules.” Online: partly.

G4. The Hollywood sequel formula. “The Martian” → “Project Hail Mary” on the “Alien” → “Aliens” pattern: not a planet but the stars, not one man but a civilization. The film rights were sold before the book came out, and the book itself is a set of scenes for the screen. Online: partly.

G5. “The engineers built in a margin” is an amateur’s view of engineering. In engineering a margin is calculated from uncertainties; it doesn’t explain away extra millions of kilograms of fuel (to M7, H30). Online: no.

G6. The paper “The Goldilocks Zone Is for Idiots” would not have passed review. Reviewers would have cut the section where Grace scolds “dozens of eminent scientists” by name (p. 46). The paper portrays a combative person whom it would cost too much to recruit (to S67). Online: partly.

G7. Grace is a traumatized character. He calls Lokken “this idiot” (and she turns out to be right) and goes off to be a teacher to gain authority without competition. It is the same thing as in his paper. Online: yes (criticism of Grace as a character).

G8. “The Thin Red Line” and the surname “Petrova”. The title is pompous and inaccurate: 26 µm is far IR, not red, and at Venus the line is as wide as the planet. It was fitted to the amnesia device (Grace recalls the letter through the “red line” of blood). The surname is the most stock Russian one; for Pulkovo, founded by Struve, a German one would be more natural. Online: partly.

G9. Cartoonish Russians. “Dr. Komorov,” “Dimitri” instead of Komarov and Dmitry; vodka, “Tsar Bomba… Boom,” “Are all Russians crazy? — Yes.” Online: partly (archetypal characters).

G10. “Elite panic” presented as fact. Stratt is sure that famine will make the US go to war for food (pp. 469–470), and the text doesn’t object. The sociology of disasters (Tierney, Solnit, Bregman) says the opposite: this is a projection of those in power that the book passes off as a property of the world. Online: yes.

Trivia

T1. “Xenon doesn’t form bonds”. Xenon compounds have been known since 1962 (fluorides, oxides); the scientist hero’s line is wrong.

T2. “On average 3500th by qualification”. Grace corrects Stratt, but she is right: at a frequency of 1/7000 the first carrier in the ranking is on average 7000th (median ≈4850). The scientist gets elementary probability wrong.

T3. Centrifuge: 3.5 rpm. This is more than the comfortable ~2 rpm: when you turn your head, Coriolis causes dizziness and nausea, and Grace doesn’t notice it. The centrifuge numbers themselves are consistent.

T4. Zylon ages. The centrifuge cables are made of Zylon (PBO), and it loses strength from light and moisture — body armor made of it was recalled in the 2000s. Spinning up during unreeling and the risk of one of the four cables snapping are not discussed (to S9).

T5. Erid’s moon is “torn away” by tides. Erid’s Hill sphere is ~660,000 km, so there is room for a moon. What would have destroyed it is tidal migration and the star’s tides.

T6. Orbital velocity at Adrian is 12.6 km/s. From the planet’s mass and radius it comes out to ≈12.2–12.3 km/s; “a little over 12” on p. 310 is correct, the figure on p. 326 is too high.

T7. Transparent xenonite in a blind species. The Eridians don’t know about light, but among the partition samples there happened to be a transparent one — exactly what a human needs.

T8. Grace teaches the blind to make light bulbs. In the epilogue (p. 514) — to a species that has no need for light.

T9. Immunity “for any crime” and waiting for trial. Stratt is protected from prosecution “for any crime anywhere on Earth” (pp. 210–211), but expects prison after launch (p. 264). This adds up only if the immunity ends with her powers, and the book doesn’t say so.

T10. A kilogram of heroin. Ilyukhina’s 1 kg (p. 461) is tens of thousands of doses, with a mass where every kilogram costs ~20 kg of Astrophage.

T11. “Nothing can quantum-tunnel through it”. Tunneling has nothing to do with a shielding layer absorbing particles.

T12. Silicon in the target. “Metallic silicon” is a semiconductor, partly transparent in the IR; graphite or copper would be more logical. And “melted a metric ton” is impossible: 1 t needs ~2.9 GJ, while 1.8 GJ was released.

T13. An AA battery for “100,000 years”. 8–50 g of Astrophage would power a house for 10⁷–10⁸ years; Hatch’s figure is understated 200–1000 times.

T14. Redell’s solar tower. A 1 km² mirror field is a full-scale plant on the scale of Crescent Dunes, not a “pilot,” and heliostats take tens of seconds to aim, not “in an instant.”

T15. Leclerc’s models and forecast. “The models have been confirmed for 20 years” — there was nothing to check them against: a 10% drop in luminosity had never happened before. The famine forecast doesn’t account for feed and biofuel (~⅓ of the harvest), reserves, and greenhouses running on Astrophage.

T16. Mummification without rot. At a humidity normal for the living Grace, the bodies would have rotted next to him for years. A ~15 kg mummy at 1.5 g weighs like 22 kg, so “weighs almost nothing” is an exaggeration.

T17. The smell of ammonia in the tunnel before the gifts. Watering eyes need ~100 ppm, ~3 g of ammonia — orders of magnitude more than would settle on the walls. Either the xenonite is porous or a seam leaks.

T18. “There are fewer magnetic storms because Astrophage steals energy”. Astrophage intercepts light that has already been emitted and doesn’t change the Sun’s magnetic activity.

T19. Cosmic rays are “only protons”. ~12% of GCRs are helium nuclei and ~1% heavy nuclei, and they account for a noticeable share of the dose.

T20. “Spinning us like a top”. Since Grace is pressed against the nose, the ship is tumbling, not spinning like a top.

T21. 10 kg of nitrogen don’t give 0.33 atm. In a 125 m³ compartment (p. 122) it comes out to ≈0.07 atm, not “back to 0.33” (p. 481). It still kills Taumoeba; the figure is what’s wrong (to H55, H56).

T22. “Retrograde amnesia… weeks”. The phrase can be read as “for weeks” or “the last weeks are erased”; in the second case Grace would remember his name (to S86).

T23. “My ship isn’t made of xenonite”. Wrong: Grace himself cut disks out of the xenonite wall of the dormitory, and the ship has Rocky’s tunnels (p. 489). The “nitrogen wash” doesn’t clean the wall — the strain hides in it from the nitrogen.

T24. The fueled Hail Mary spends 16 years in synchronous orbit around Erid. The orbit lies in the inner radiation belt, and a ship with leftover Astrophage hangs over an inhabited planet — a repeat of S78.

T25. Errors in the Russian translation. “A few days” instead of “a few years” (fuel accumulation), “protons” instead of “photons,” 245 bombs instead of 241, the cyclone season shifted by “two weeks” instead of two months.

T26. The Majorana nature of neutrinos is presented as fact. Neutrino–neutrino annihilation, on which the Astrophage battery is built, relies on the neutrino being a Majorana particle. This has not been established experimentally, yet the book presents it as a known fact (see H14).

T27. Blackpanel: one sheet of glass, not “on both sides”. The original says “Put clear glass over it”: glass on top only; “on both sides” is an error of the Russian translation. The estimate of glass mass for the farm therefore remains 6–23 Gt (S14, see T25).

T28. Tau Ceti e most likely doesn’t exist. Adrian corresponds to the hypothesized planet Tau Ceti e. According to a NASA review, studies in 2024–2025 showed that the signals of the four planets announced in 2017 are more likely explained by stellar activity or data processing; the novel’s planetary system is a fictional model.

T29. The external cameras and antennas at Adrian should burn out first. Exhaust reflected from the atmosphere delivers ≈10⁵–10⁶ W/m² (100–1000 “suns”) to the hull. The skin is at least somewhat cooled by Astrophage, but external cameras, antennas and sensors are not, yet in the book only one camera goes out (see H45).

T30. “High gravity… bones degenerate faster” is the reverse of the truth. In the epilogue Grace explains his weakness by saying that bones degrade faster under high gravity. In reality load makes bone denser, and a person loses it in weightlessness; at 2.1 g the joints, heart and veins suffer and the risk of falls increases, so arthritis and a cane are plausible, but the explanation is wrong (see S108).

Numbers S29, S44 and S87 are not used: those items were merged with others or moved to holes.

Found by Others, Missed by Me

ComplaintNote
The tanks are too small for ~2 million kg of Astrophage: per the ship’s schematic they’d need to be 2.5–3 times longernot checked; see H11 (paste density)
The direction of g-load in ch. 19 is described at an angle to the “floor”minor (moviemistakes)
Jettisoning through the airlock at 10% pressure won’t push a body out at 1.5 gclose to H43, but different mechanics
The Sun’s rotation period from sunspots near the limb is nonlinear (early chapters)disputed in the thread
The “cube root of 8” scene: the computer rejects the correct answerminor
Mitochondria in a cell in space with no electron acceptor (O₂)see H36, H37
Eridians sleep under guard: Rocky’s 46 years of solitude are psychologically impossibleWatts; see S59
It would be smarter to run the Taumoeba selection at home rather than on the shipI accepted the heroes’ decision
The mission has no standard sampling equipment (probe, shuttle)partly see S76, S99
The computer doesn’t warn about an approaching alien shipsee S66
Grace’s name on the patch and robot recognition despite the swap days before launchKowal (WaPo); fans’ answer — Stratt was preparing him in advance (see S103)
Up/down after the ship flips (ch. 20–22); Rocky says “black” before having a concept of colorminor
Walking in a ~120 kg Orlan at 1.4 g is implausiblesee S108