The best thing I read this week is Ted Chiang’s piece in The Atlantic, No, Artificial Intelligence Is Not Conscious.

Naturally, I have to start with how cool it is to read one of the best living sci-fi writers (best known for Story Of Your Life, the basis for the film The Arrival) writing about something I talk about on coffee breaks.

He raises a few arguments that matter for a lot of questions, not just whether AI is conscious.

Here’s the core one, in short. If you showed me a video tomorrow of a spacecraft orbiting Alpha Centauri (4.3 light-years from here), what in that video would convince me it was real? Nothing. First we’d have to see astronauts orbiting Mars, then Jupiter, then the rings of Saturn… It’s very hard to believe anyone who claims they solved an enormous problem without seeing how they solved the easier ones that come before it.

This connects, in a way I find interesting, to a very different trend: behind-the-scenes of TV shows, fashion shoots, influencers. On one side they show you the perfect, polished product (remind you of anything?). On the other they show you the chaos, the mess, the dirt behind the production. We’re not only looking for authenticity. The path and the process aren’t separate from the final product.

A lot of the time the discussion about AI starts at the final product. That makes sense, it’s the academic method (let’s assume X, and see what it means for the world). The problem is the discussion also stops there.

Want to do a degree? Forget it, it’ll replace us all. Plan a sandbox? No point, it’ll break out of everything. Start a company? Write a book? Reach space? Optimize a GPU? Don’t bother, matrices will do it better than you.

Some kind of god-mode you can’t argue with. The discussion is over before it started. When you’re up against an entity that solves everything, every practical question sounds petty. And that’s not just a scary description, it’s a perfect excuse not to work. If it’s going to solve or break everything anyway, why bother with the boring problems in the middle?

Except… anyone who actually works with this technology knows there’s a big gap between scoring on benchmarks and the real world. The road ahead is long and packed with technical challenges we have to solve, with products that haven’t been built yet, with human and commercial problems that I find exciting. I’m not impressed by anyone who tells me they’re not worth the effort.

I’m not saying we won’t reach Alpha Centauri. I just want to see Saturn on the way.