michael-dean-k/

SPOILERS: the movie starts with Ryland Grace (Ryan Gosling) waking up in space from a coma; he has amnesia, and as we watch him learn that everyone else on the ship died, we're hit with the dread of wondering how this untrained school teacher is going to operate an interstellar spaceship. Naturally, I imagine myself in that situation: I'd panic to find manuals and properly labeled buttons before accepting my own death. Even if I could figure out how to steer a ship, I could never land it. This whole daydream collapsed when you realize the ship controls itself. It's automated.

In the process of Grace troubleshooting how Tau Ceti, a distant sun, was able to resist astrophage (a cosmic sun-eating plague), he faces a chain of similarly impossible obstacles, including a friendly alien who he cannot communicate with. The stakes are enormous. If they can't cooperate to solve this problem, both species go extinct. In a matter of days, Grace and Rocky (the rock alien) are able to build translators and speak fluently through their technology. Compare this to Arrival, where it took months to achieve this, or how it took years for Europeans and American Indians to smoothly communicate. We have to imagine that Grace, a teacher and world-class biologist, is not an expert in linguistics, nor a computer programmer, and so he certainly couldn't have built this—an AI on his computer did.

As I watched it I wasn't at all thinking through technical details or questioning Andy Weir's science. I surrendered to the world, as you should. I'm sure the science is all well-researched and convincing. As a viewer, I suspend disbelief and take it in at the layer of emotion: the unlikely and charming relationship, plus their willingness to risk their species to save a friend, is all very moving.

But something that sits with me is the absurd competence of a single man, Ryland Grace. He was a school teacher hiding from a controversial biology paper he wrote, and ended up being the one to discover how astrophage reproduces (leading to its use in fuel for interstellar travel). That's already a massive species-level accomplishment. But after most of the trained crew died during an explosion during training, the project lead (Stratt) asks Grace, who is not trained as an astronaut at all, to go on a one-way suicide mission. When he refuses, they detain him, put him in a coma, and put him on the ship. Against all odds, he makes even more species-level breakthroughs to save the species. It's as if this humble middle school teacher is actually a space age da Vinci, making breakthroughs in biology, chemistry, physics, programming, and linguistics, all in a couple of weeks, all while recovering from amnesia.

Another way to think of this: Grace is an expert in one field, biology, but when equipped with a powerful AI, he's able to generalize his thinking and make breakthroughs in every field. That makes it all plausible.

But the story has no explicit mention of AI. Weir was intentional about this. There are medical robots on the ship, but are purely mono-functional. The ship operates through AI, but there is no HAL-like personality. The figure of the externalized genius takes the form of the alien, Rocky, the master toolmaker. Worth noting that this story was written 2018-2020, before mass awareness of the rise of generalized AI, and so the interstellar ship really runs more like an advanced Siri.

If we were to experience a real-life astrophage crisis in the 2030s, AI would be central to every facet of our response. The project team only had 3 years to invent a new type of fuel, an interstellar ship, automated coma-induction technology, a self-navigating ship, etc. That sounds like 30 years of progress in 3 years, the exact kind of technological compression everyone is forecasting with AGI.

This story exists in a weird and unlikely sliver where we can ship out mankind for interstellar travel, but we don't yet have autonomous AGI that would automate the dangerous and sometimes impossible jobs of astronauts.

Part of writing science fiction is properly imagining future technology stacks so that the film can appeal to both the emotional and analytical mind. When you get the world wrong, it creates unreal situations. The main example for me: the interstellar trip was a one-way mission—if they find the cure to heal their sun, they're instructed to send it back through four autonomous interstellar space probes (named "Beetles," each labeled "John," "Paul," "George," "Ringo"). If they can send probes from Tau Ceti, why not also send them to Tau Ceti and prevent the whole suicide mission? Pre-AGI, the answer is, well, sure, you can send probes to a foreign solar system, but once they get there, they won't be able to do science, they won't be equipped with the supreme problem-solving skills of Grace and Rocky. But post-AGI, those probes would be way more equipped than Grace, and since they aren't biological, they have no fuel requirements (powered by solar), can travel way faster, and are much better suited for the hostilities of space.

Space is a supremely inhospitable landscape, way more ferocious than the oceans of the Middle Ages, and once we have AGI, it becomes a lot less likely that we'll send human astronauts into space to do science. If we had a real-life astrophage crisis, we would not send sacrificial astronauts; we would send millions of autonomous probes to Tau Ceti, spaced out to created a real-time communication stream back to Earth, each specializing and approaching the problem from different angles.

Of course though, the scientifically accurate story is lame, brittle, and lifeless compared to the heroic astronaut who overcomes his cowardice and becomes best friend with a rock, retiring to his alien planet. Unliked hard sci-fi viewers, who are sticklers for realism, I prefer the human drama. But it seems like a lot of science fiction starts with a simple conflict, a high-concept world, but then styrofoam details that the majority of viewers will never notice. Weir is not like this. Weir is considered to be a hard science fiction writer, but I wonder if AI throws a wrench in forecasting. It's so totalizing in how it touches every branch, making all branches unknowable in the near future. Can we not having a gripping human drama in a properly forecasted AI future? Or does totalizing automation kill the circumstances for drama to arise?