everyoneShouldUseGit redd.it/1g6d83m

Monte Carlo is unnecessary. This has an exact closed form solution

pip error: numpy must be installed in a superposition of different versions in order to satisfy all mutually contradictory version requirements

diffusion models are just a ploy by CS PhDs to get their departments to pay for them to learn stochastic calculus so that they can get finance jobs

llama uses rmsnorm, r1 uses rmsnorm, grok uses rmsnorm, o1/o3 use [redacted] so why doesn't pytorch have a native (fused) rmsnorm in the year of our lord 2025? not for lack of user requests e.g., github.com/pytorch/pytorc… 1/n

never ask a man these two things: - his income - if his grad norms are indicating something

"Listen we are actually using a powerful AI model"

due to tariffs the optimal learning rate has dropped from 3e-4 to 5e-6

i can't believe the last stage of using linux is using macos

The least beautiful equation in mathematics.

im actually so bad at EQing... i got low-shelf esteem...

We made an MCP Server that can extract system prompts from agents 🤠 Evan Fabry

the lion does not concern himself with revenue

Don’t fall for big saddle point propaganda

New paper alert! We report that the Hessian of NNs has a very special structure: 1. it appears to be a "block-diagonal-block-circulant" matrix at initialization; 2. then it quickly evolves into a "near-block-diagonal" matrix along training. We then theoretically reveal two

f’’’’’’’(x) gang

At this rate Google is gonna drop GTA VI before rockstar

Confirmed: 100% AI generated Sam Altman