It’s amazing in this day and age I have to tell you idiots to stop tweeting about the product you hate.

I volunteer to edit/review for a Silicon Valley Review of Books, in case anyone is bored/interested in a doomed-to-fail venture.

Idk how to phrase this right but if you see a cool release and say something like “they must have had an algorithmic breakthrough” then I think your mental model of the world is full of type errors.

I think a lot of the LLM naysayers are primed to think a Move 37 analogy is impossible in the current AI stack, similar to the pre-AG skepticism we were anywhere close. Luckily, like AG, when it comes the skepticism will disappear fast.

But what would this look like? If someone asked Chat to solve the Riemann hypothesis and it did, that would clearly qualify. But if it was a prolonged convo between humans and AI? I imagine the skeptics would (somehow) be unconvinced.

Sam & team are exceptional and we’re looking for more exceptional people! The London team is phenomenal :) if you’re excited to make the best models in the world even better, come join us!

This is a pretty AI-centric account and normally I don’t engage in other things but for the record this is one of my favorite scenes from one of my favorite movies. I died laughing in the theater.

I want a feed for high-perplexity takes only

Accidentally said “where to next?” instead of “good night” and they kicked me out of SF

See book by Herzog at new non fiction shelf at SFO. Flip through a couple pages, see the word “AlphaZero”. Sold.

I had something insightful I wanted to say about architecture research … but I forget what it was. I have it written up at home somewhere!

This is really true about OAI but I imagine it’s mostly true at a lot of places and you can usually tell (and should disregard the takes of) someone who has never worked at a large company assuming the actions of the org are those of the leader.

Some free advertising: Excalidraw is great!