🔥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

Today, we at OpenAI achieved a milestone that many considered years away: gold medal-level performance on the 2025 IMO with a general reasoning LLM—under the same time limits as humans, without tools. As remarkable as that sounds, it’s even more significant than the headline 🧵

all languages should be gpu languages!

One Happy Fellow Here's my attempt at making programming language semantics more widely accessible: tinyurl.com/semantics123. One day I'd like to write a book about this.

so chess, go, and icpc/imo is done. the next frontier/update is - pointing an llm at an open problem (openproblemgarden.org) and ask to solve - pointing an llm at linux/clang-llvm/sqlite code base and ask to cut a pr to increase perf by X%

know your alignment, padding, and sizes folks 😎

finally have my head wrapped around scalar compilation 😎 for j4orz.ai/zero-to-hero/, and now giving myself permission to dive into gpus. singularity systems: zero to hero is going be following the gpucc paper from cgo '16 dl.acm.org/doi/pdf/10.114… by extending the big red