People find out I read twenty-three AI newsletters (and unsubscribed from as many more) and immediately ask the same question: which ones should I read?
Fair. Here’s what’s in my rotation and why.
First, the infrastructure. Beyond newsletters, I have a custom scraper that pulls new AI and HCI papers published each day. I built it myself—by asking Claude. Teresa Torres gave me the idea, so I went and asked the AI how to make my own. I was even able to add a simple ranking algorithm and have it generate plain-language summaries so I can quickly tell which I want to read. You can do the same. Building stuff is better than reading stuff for learning GenAI. Do both and you’ll be caught up in no time.
I also listen to some podcasts on AI and follow a bunch of folks on LinkedIn who talk about AI. I’m not on Twitter anymore (Damn you mecha-hitler!) and not on Facebook. I rarely pay for newsletters because I subscribe to so many and I’m on a lecturer’s salary.
Why do I do this to myself, you ask? I am a very fast reader. I am a very good pattern matcher. And I’ve lived through three massive tech revolutions and I know what they look like. I don’t think everything will change with AI but my job sure will.
If you want to know what’s happening
Superhuman by Zain Kahn is my go-to daily driver. I especially love that he does robotics news on Saturday and nifty science on Sunday. AI Breakfast is another good one, I like their long laundry list of articles. And if you prefer audio, the AI Daily Briefing podcast is my dog-walk companion—his endless optimism makes me happy. Okay, he’s a hypester and he does have an AI business, but he also has real enthusiasm. YMMV.
If you want to understand what it means
Benedict Evans is required reading—smart, skeptical, long view on tech. Ethan Mollick (One Useful Thing) does deep dives on specific topics and is running some of the most interesting public experiments right now—a musical featuring AI-generated otters being a recent highlight. Follow him on whatever social platform you prefer. His blog is great for deep dives, but if you don’t follow him on the socials, you’ll miss his sillier moments which are delightful.
If you want to stay conversant in the tech but you’re not a programmer
Teresa Torres and Ben’s Bites both claim to be non-technical. Don’t believe them—they’ll keep you genuinely up to speed on what’s happening under the hood and how to wrangle it with agentic coding. LLM Watch is a gem; their “Papers You Should Know About” keeps me grounded in what models actually can and can’t do. Decoding AI does deep dives on core concepts like evals—useful when you spot a term in the daily news and want to actually understand it. Paul Ford on LinkedInis practically a newsletter, he’s constantly posts great finds.
One honest caveat: I always read Nate’s Substack and the insights are real. But the writing style is… let’s say it has a certain AI-generated stank. GenAI-brained readers will be fine. If you care about a lovely sentence you may not survive.
I’m not a programmer and a lot of stuff flies over my head, but I do like Simon Willison’s blog.
If you care about writing
every.to is required reading—they actually care about prose, which is rarer than it should be. Mac users, it’s worth paying for; the subscription includes some well-regarded software. (I wouldn’t know personally. Like John Hodgman, I’m a PC.)
I also just stated reading Ben Recht (arg min) — ML research skeptic — and his writing is delish. Check out his series on vitamins!
If you’re a PM
Lenny’s Newsletter has gone very AI lately, which also gets you How I AI—hit or miss for me, but your mileage may vary. Teresa Torres is getting so technical lately she’ll soon be best for engineers, but she keeps me conversant in the details. Aakash Gupta has been doing the growth newsletter game for years before getting thoroughly GenAI-pilled, and he genuinely understands where AI fits into the PM job rather than just stapling “AI” onto existing frameworks. I can’t remember if he led me to Tal Raviv or vice versa, but I like them both and you should follow both.
If you’re a designer, the only newsletter writer I find a ton of value from so far is Xinran Ma. He’s working hard to figure out what does and doesn’t work for design. Suggest some?
There are so many UX folks I follow on Linkedin that have good things to say, and the ones I forget to list will yell at me, but I’m trying to make a shortlist so try out Jorge Arango, Dan Saffer and Sam Ladner. Sam is an amazing design researcher and a very reluctant adopter of AI and her advice is gold.
For when you’re getting too bullish
This is important. Unrelenting AI optimism will rot your brain. I follow Timnit Gebru on LinkedIn—she leads the DAIR Institute and will de-hype you fast. Karen Hao, who wrote Empire of AI about OpenAI’s rise, just started a podcast and is worth following closely. And Emily Bender—of “stochastic parrots” fame—is essential for keeping your epistemics clean. None of them are anti-AI for sport. They’re just rigorous, which is exactly what you need when everyone around you is drinking the Kool-Aid.
A note on Video
I don’t follow AI YouTubers regularly, but when I was first trying to understand transformers a couple years ago, I watched a few to get my bearings. IBM had some great primers. Worth doing. Understanding the tech at even a high level matters—you’re far less likely to mistake the model for a sentient being.
Should I post more from my reading, or do I already yap about AI too much? Tell me honestly. Should I start up another newsletter, a meta-newsletter of newsletters?
Written with, but not by, Claude.ai. Full list of newsletters here