Excited to share that I am starting a new company dedicated to the creation of verified superintelligence via autoformalization. Building on the amazing RL infrastructure that we have developed at Morph, Math, Inc. has already achieved a breakthrough result by

Congratulations to the Math, Inc. team on their announcement! AI-assisted autoformalization is a promising area with significant potential impact. We're looking forward to the community discussion and seeing how this space evolves. #AI #ProofAutomation #FormalMathematics

In the first year of my PhD, I formalized the independence of the continuum hypothesis in Lean. It was considered a breakthrough at the time, taking ~12 months and 20K lines of code. With Gauss, we finished Strong PNT with 25K LOC in 3 weeks. There are two really remarkable

I tend to agree. One caveat - will people who don't need glasses, be willing to wear them all the time?

AI has officially beaten me at the ICPC World Finals. It reminds me of a rare ICPC skill: being able to quickly read a teammate’s code and spot bugs. This skill takes years to train, and explains why AI often makes coding slower (see arXiv:2507.09089). No matter how strong AI

I gave a demo and talk last Thursday. It is about a self-sustaining programming environment based on Functional Reactive Programming. In my mind, it is not a mere hand waving demo; the environment is capable of creating non-trivial size applications. youtube.com/watch?v=fRX6Xq…

Software should be soft: malleable, adaptable, and shaped around your brain (not the other way around). Few have written about this better than Geoffrey Litt. SO thrilled he joined Notion last week! If you get a chance, do read this essay: inkandswitch.com/essay/malleabl…

Finally had a chance to listen through this pod with Sutton, which was interesting and amusing. As background, Sutton's "The Bitter Lesson" has become a bit of biblical text in frontier LLM circles. Researchers routinely talk about and ask whether this or that approach or idea

I had a great discussion with Geoffrey Litt, Nintendo .DS_Store and Max Schoening after hours at Notion. We talked about using AI to augment intellect and augment tool-making. I learnt a lot! x.com/akothari/statu…

Nice. Live programming, snapshot and time traveling debugging in a well integrated package. This could and should have been the norm decades ago.

Paul Snively Also, there is a school of computer "science" I utterly despise according to which a programming language semantics is the same as static types. Nope. The real world, the extent programs and programming practices, come first, and if your "science" is unable to analyze it, well,

I just uploaded videos for the first lab of my tiny systems course. TinyML - Write your own tiny functional language interpreter! 🎞️Watch the videos on YouTube: youtube.com/playlist?list=… ⌨️Get the source code here: github.com/tpetricek/tiny…

AI is at the stage of transportation during the steam-engine railroad era: clunky, expensive, unreliable, and inflexible. However, diesel engines, cars, airplanes, and rockets will soon emerge, with progress now 30 times faster than then.

Hearing from a lot of good founders that AI tools are writing most of their code now. Software engineers orchestrate the AI. They are also finding it extremely hard to hire because most experienced engineers have their heads in the sand and refuse to learn the latest tools.

Wondering if the folks who think "you need static typechecking to create high-quality software" experience any cognitive dissonance as they type their program into Emacs?

The video recording of my ICFP 2025 keynote talk Functional Programming for Hardware Design is up on YouTube. youtube.com/watch?v=1oBOu6…

What's often annoying about debating static type aficionados is that they argue as if static typechecking, unlike any other tool in existence, didn't come with a set of trade-offs.

If people like this would spend a couple of days playing with classic, decades-old dynamically-typed systems like Common Lisp, Smalltalk, and Scheme, they would see that they are superior in some important dimensions to anything else out there. But no, that won't happen.