#Phrack is much older than me, it is good to see it still doing so well, and an honour to be one of its authors in the 40th anniversary edition, which is now available online: phrack.org/issues/72/9_md… Thank you for what you are doing for our community. Long live the #Phrack!

an unexpected result of incredibly good AI agents is that you start to study more and read more it's because your own knowledge becomes the only differentiator, when everyone else has an access to the same models and know the same prompting tricks as you

Some of my recent PyTorch compiler contributions are now released in PyTorch 2.10, especially take a look at dp_knapsack_sliding_hirschberg which might speed up your default PyTorch a bit and also allow you to run bigger models 🫡 pytorch.org/blog/pytorch-2… github.com/pytorch/pytorc…

It's really bizarre that the premier AGI lab on the planet (outside GDM, who are falling behind in actual products again) is so heavily Polish. 6 top OpenAI contributors have been Poles. Imagine what Poland + more GPUs could have done. Would they mog China? The US?

Stanislaw Lem in 1986: „The first inventors of machines that augmented not the power of muscle but the power of thought fell victim to a delusion that attracted some and frightened others: that they were entering upon a path of such amplification of intelligence in nonliving

限られた人間しかプレイできないと言われているノイズテトリス

💤Researchers found that prescription stimulants for ADHD act on brain networks that control wakefulness and reward, but not attention as previously thought. The study suggests that stimulants and additional sleep affect the brain in similar ways, and that getting enough sleep

One not very hot take - The Claude C Compiler has the best internal architecture docs of any compiler I've ever seen. Far, far, better than any compiler I've ever written, lol :-)

will CogInterp Workshop @ NeurIPS 2025 Workshop be held in NeurIPS 2026? 🐱

AI is writing a growing share of the world's software. No one is formally verifying any of it. New essay: "When AI Writes the World's Software, Who Verifies It?" leodemoura.github.io/blog/2026/02/2…

www-cs-faculty.stanford.edu/~knuth/papers/…

How are mathematicians facing the wave of rapidly advancing AI-for-math capabilities? Jeremy Avigad (CMU prof and co-author on the original 2015 system description paper for Lean) just posted a paper with his thoughts in the wake of the Math, Inc. announcement on sphere packing.

The FA4 paper is finally out after a year of work. On Blackwell GPUs, attention now goes about as fast as matmul even though the bottlenecks are so different! Tensor cores are now crazy fast that attn fwd is bottlenecked by exponential, and attn bwd is bottlenecked by shared