The Reverse Centaur's Guide to Life AFTER AI

I can't fault Cory Doctorow for anything, really. It might be that I'm biased (because he wrote me back when I did one of his books for an assignment in Grade 9), but his latest book truly offers us such an insightful and level-headed perspective on the developments in AI, or perhaps I should call them LLMs.

It's cliche, but AI is brought up everywhere now. From convocations (where it is summarily booed) to corporate meetings (where the boos are silent), AI is talked about so constantly that it's kinda annoying. Maybe that's why I put off reading this book: I was fed up with listening to the endless debate between shills and critics.

Cory is not so reductive. He's too smart for that. There's a well justified cynicism in the way he talks about these AI megacorps, but it's not the usual arguments of techno-pessimism. The poor understanding of the technology is the common thread in all the discourse that seems to annoy him more than anything, but what genuinely concerns him is not the existence of LLMs as such. It's this sort of maxim he has that repeats throughout the book and (spoilers!) is repeated at the end:

"The most important fact about a technology is not what it does, it's who it does it for, and who it does it to."

Well said. What I most enjoy in looking back on works about technology over the past 100 years or so is precisely the techno-optimism that is sorely missing from conversations about technology now. For instance, it is another cliche to bemoan how we are all glued to our phones, addicted to social media, or have our attention spans eroded by a fatal diet of viral shorts. But what's actually so bad about phones, or staying connected, or having a laugh at something funny? Nothing.

It wasn't even that long ago that calls were exorbitantly expensive if you dialed long-distance, never mind another country! You would have to spend hours researching something that could be accomplished today with a quick search on Wikipedia. Oh, and do you remember using a typewriter to create important documents? One mistake, you'd have to redo the page or mark the mistake with more X's than a porn site. These are the promises of technology, and it's truly remarkable what it has given us in a relatively short span of time.

So what's the issue? Doctorow nails it: "who it does it for, and who it does it to."

Your emails aren't free because Google has an altruistic mission.

Social media isn't free because Zuck wanted you to "poke" your estranged family members.

These aren't products for YOU. YOU are the product for someone else, a captive audience. For all the good technology does, it will do what you tell it. If I told my computer to blow up a truck of orphaned kittens, it would do it if it could. (Currently, it just says: "bash: blow-up: command not found.") It does these things because someone tells it to do them. So, if you're not happy with AI or anything else for that matter, best look to the people who are responsible. It's not technology that's making you miserable. It's people.

The image of a reverse centaur is a good one. Centaurs, the mythic half-human half-horse hybrids, make good use of horse legs for getting around and what not. They're badass. Intelligent. Quick on their feet--erm, hooves. A reverse centaur, however, is not found in myth, and for good reason: there is no fathomable advantage to having a horse head and human legs, unless you want to act in a rendition of Midsummer Night's Dream. With technology, that would be like having PowerPoint do the presentation while you shyly advance the slides in front of everyone. How embarrassing! I'd much rather be hung like a horse.

A centaur isn't a centaur if it's a reverse centaur. That's just a gag from Shakespeare. In the same way, technological progress might not equate to actual progress. After all, you can have great advancements in technology that are catastrophic for life, the universe, and everything. We see this quite clearly in war-time. Alan Turing, who previously developed a general theory of computation, helped crack German Enigma codes for the Allies, arguably shortening the war and definitely saving lives. On the other hand, the ENIAC computer's first major calculation was for the H-bomb. Same field. Same decade. Very different outcome for humanity.

So ask yourself: do we exist for technology, or does it exist for us? (Sci-Fi sideline: maybe if we ever do achieve AGI, both can be true?) If you think it should be the latter, you're a centaur; if the former, you're a sucker.

I'd encourage you to use technology in an empowering way, however you want to imagine that. That's true even of AI. I'm sure you've had at least one problem it's solved for you. You can even download a local model right to your PC! And there are programmers, engineers, and projects that do a lot of good in AI and beyond. They do it for free (or donations), mind you; you just won't read about them in the Wall Street Journal. But you might hear about them from Cory.