Dispatches from Cydonia (or: Who's worth reading about AI)
Hi there!
It’s been a while. I took a 2.5-year pandemic-motivated detour in my career (future newsletter topic), and since I wasn’t doing any public writing, I didn’t do any newsletters either. But now that “AI” is causing the latest Great Derangement of our times (and I’ve started to write about it again), I figured it’s a good time to chime back in. With an elaborate nerdy metaphor, as is my wont.
Cydonia is the region on Mars where this 👆 infamous face was spotted in 1976. The face is, of course, an illusion—but a perfectly captivating one. Even if you don’t believe it’s “really” there, you simply can’t not-see the face.
You see where I’m going with this. The public release of ChatGPT was AI’s “face on Mars” moment—and as a result, the whole topic of AI has become Cydonia. Everything in the field suddenly seems to be captivated by the supposed “intelligence” at work in this region; all the subsequent reporting, doomerism, anti-doomerism, anti-anti-doomerism, etc. is emanating out of it, too.
But instead of dropping a 12,000-word personal take on you, for now I just want to share my shortlist of who's worth paying attention to about all this. These dispatches from Cydonia help me keep the actual terrain of AI visible, while providing lucid (but arm’s-length) commentary on all the face-chasing.
1. The Reporter: Tiernan Ray, ZDNET
My man Tiernan does not do hot takes. He’s very nuts-and-bolts, and sometimes too hardware-focused for my taste. But here’s why I keep coming back: he’s like the Columbo of AI. Unpretentious, unassuming, and utterly unbullshittable. He’s seen it all and finds things out that I’ve never heard of. His writing has (if you’ll indulge another Peter Falk reference) a “Grandpa in The Princess Bride” vibe: affable but authoritative, with a dry gruffness when something tries his patience. (While reading his response to Kevin Roose’s “Bing tried to break up my marriage!” freakout, I could almost hear Falk quipping to Fred Savage: Yes, you’re very smart. Shaddup.)
2. The Expert: Melanie Mitchell, Santa Fe Institute
This won’t surprise many of you, given that she wrote a very well-received primer on AI last year and is active on Twitter. What makes her indispensable to me (vs. other scientists I agree with, but rarely read) is her coolness: she has the expertise to mix it up with AGI hypebeasts and credulous LLM researchers, but she hasn’t turned "losing her shit over their nonsense" into a brand. She calls balls and strikes in her newsletter like a scientific umpire, and also surfaces challenging new findings. There’s no exasperation or sourness to her AI commentary; she was a protégé of Douglas “Gödel Escher Bach” Hofstadter, and it shows in her deep and genuine curiosity about what this tech can actually do.
3. The Stranger: Jan Chipchase, Studio D
This guy was once known as the Indiana Jones of design research. He doesn’t have much to say about how the tech or science of AI works, but he has tons to share about what people believe and do about it. His semi-regular newsletter technically isn’t AI-focused, but the most recent one was full of unusual, fascinating links on the subject—and I expect more of the same in the near future. His dispassionate, “anthropologist on Mars” POV is a bracing corrective to cookie-cutter takes on AI (especially the ones I happen to agree with).
4. The Others
Folks I read less frequently, for varying reasons.
- Frank Lantz, a game designer/artsy weirdo whose new newsletter is witty, thoughtful and genuinely surprising. I’m not sure yet if I’m in for the long haul, but this post definitely hooked me. (TLDR: your arguments about AI may be airtight, but if your thought experiments are tacky? Dealbreaker!)
- Jack Clark, the OG AI newsletterer. I’ve been reading him less since he went full Locutus-of-Borg as a cofounder of Anthropic (aka the “Trust Us This Time—No, Really—Why Are You Laughing” AI company), but I can’t bring myself to unsubscribe. He still has the best tech-trend updates.
- David Chapman, an old-school AI researcher who wrote the single best one-shot 101/“what it all means”/master-metaphor-for-the-whole-thing I’ve ever read. His recent anti-AI e-book is kind of a mess (thanks for that, Frank Lantz), but that one essay was so laser-bright that I’ll sit up straight for anything new he has to say.
5. [ _____ ]
You made it to the end! What’s your favorite dispatch?
Here's what I'm looking for and haven't found yet: a George Orwell of AI. I want a "Politics and the English Language" for multimodal transformers. A "Shooting an Elephant" for AI safety. A Homage to Catalonia, but for Cydonia.
Know anyone?
That's all for now,
John
