How to become expert at thing: 1 iteratively take on concrete projects and accomplish them depth wise, learning “on demand” (ie don’t learn bottom up breadth wise) 2 teach/summarize everything you learn in your own words 3 only compare yourself to younger you, never to others

sarv Inspiring and true but incomplete. It should be augmented with recognising the possibility of failure. To be joyful at that possibility, one has to find ways to enjoy the quest itself. Then one can be joyful if one ends up filling in no blanks, discovering only a new way to fail.

Trying to understand vs. trying to build are actually two different things. Sometimes they synergise, sometimes they are in conflict. Often they are orthogonal. They direct attention differently. No need to identify with either though.

"I think the best way to prepare for the future 20 years is find something you love to do, to have a shot at being one of the best people in the world at it. Build an independent brand around it, with your name." Naval

"If you define jobs in terms of tasks maybe you're actually defining away the most nuanced and hardest-to-automate aspects of jobs, which are at the boundaries between tasks." 👏

"Your goal in life is to find the people, business, project, or art that needs you the most. There is something out there just for you." Naval

Yes. A few miscellaneous thoughts. (1) First, the new bottleneck on AI is prompting and verifying. Since AI does tasks middle-to-middle, not end-to-end. So business spend migrates towards the edges of prompting and verifying, even as AI speeds up the middle. (2) Second, AI

As a 31yo Principal Engineer with 20 years of programming experience and 15 years in the industry, I I would add that you should know Vacuum fluctuations Planck-scale quantum foam Feynman path integrals Quantum electrodynamics Quantum chromodynamics The Standard-Model Lagrangian

Both LLMs and LRMs are upper bounded by humanity's knowledge closure. True scientific discoveries are, by definition, outside of that closure. Ergo, LLMs/LRMs are great force multipliers to us; but don't support "Nobel this weekend" hype.. (🧵 from yesterday 👇👇👇)

No book can ever be finished. While working on it we learn just enough to find it immature the moment we turn away from it. — Karl Popper

We achieved gold medal level performance on this year's IMO! Our model thinks and writes proofs in clear, plain‑English - no formal code required. Unlike the narrower systems used in past competitions, our model is built to reason broadly, far beyond contest problems.

This is not to say that finding everything in humanity's knowledge closure is trivial! Most math is indeed in the deductive closure of the axioms.. and no wonder LRMs continue to have impressive success there (modulo the curmudgeon Gödel and his eventual incompleteness.. )

Really cool.

I’ve just published a new blog post: In Defence of Discussion. open.substack.com/pub/samuelkuyp…

Still can’t believe it. Our work on uncovering plagiarism in AI generated research received the outstanding paper award at ACL!! Effort led by, and envisioned by the amazing Tarun Gupta!

Thanks for the outstanding paper award!!

Acid test from Naval to avoid fooling yourself: “Would you still do it if you couldn’t tell other people about it? If you would, it’s real. If not, it’s just a social ego boost.”

Skepticism is a soft inductive bias for good science..

Well said!