🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Instantiation in educational program surveys: every program I've been to gets about a 9/10 rating to the question meant to measure counterfactual impact. OK OK even if it wasn't "meant" to measure, it will later be used to make this claim.

"substantive" economics and mathematics discussions in academic settings are only bearable to me when i go in with the intention to make notes on how people will miscommunicate. i do this about once a week, also the frequency of my desire to start a new teaching program.

it is interesting how much search results change when googling symptoms vs symptoms + "woman"

games i have experienced strong tetris effect for so far: 1. wordle (+variants): thinking in 5 letter words 2. set: 2 colorful images popping up, find the third one 3. kakuro: quiz me on how many ways there are to make 5 unique single digit integers add up to 28

Love this little chart here from Open Phil

Are you afraid to value bet unless you’re almost positive your hand is good? It’s not uncommon, but you’re missing out on a massive amount of value and it’s costing you $$$!

Obvious but surprising fact from analysing my self-reported daily productivity metrics from 2023: using AI tools was anti-correlated with deep work. I would like to experiment with this more, but don't have any bright ideas for it.

There's no clearly winning score on OCEAN, but there is a losing score. Smh. Just like BMI.

MSJ is an order of magnitude more effective at jailbreaking Claude than SOTA attacks!!! The paper is a cool step in thinking about *model capabilities* as *attack surfaces*. Congrats to Cem Anil on knocking it out of the park! Super glad to have been a part of it!

I believe much more of our bandwidth should be spent getting ready for the next generations of models, on security, monitoring, preparedness, safety, adversarial robustness, (super)alignment, confidentiality, societal impact, and related topics.

Mechanistic interpretability gives us rich explanations of models. But can we convert these explanations into formal proofs? Surprisingly, yes! Mech interp helps write short proofs of generalization bounds — and, shorter proofs provide more mechanistic understanding. 🧵

Theorem (Theorem) is an AI-coding IDE for mission-critical software. They're making program verification 10,000 times faster, so your systems code does exactly what you asked for. Congrats on the launch, Jason Gross and Rajashree Agrawal! ycombinator.com/launches/NZA-t…

If Andrew says so, it must be true 😁

An interesting concept! I'm curious to see where it goes