LLM Neuroanatomy II: Modern LLM Hacking and Hints of a Universal Language?(dnhkng.github.io)

98 points by realberkeaslan 7 months ago | 30 comments

LLM-powered tools amplify developer capabilities rather than replacing them(matthewsinclair.com)

323 points by matthewsinclair 1 year ago | 214 comments

Blog post by Matthew Sinclair arguing that LLM-powered programming tools act like a mech suit amplifying developers rather than replacing them.

TLA+ in Isabelle/HOL(davecturner.github.io)

71 points by kalaracey 2 years ago | 2 comments

LLMs and Code Optimization(wiredream.com)

139 points by dgacmu 1 year ago | 22 comments

The changing goalposts of AGI and timelines(mlumiste.com)

333 points by skandium 7 months ago | 252 comments

Blog post arguing that OpenAI's 2018 charter self-sacrifice clause, combined with Sam Altman's stated AGI timelines, implies OpenAI should stop competing.

Task-free intelligence testing of LLMs(marble.onl)

59 points by amarble 9 months ago | 17 comments

The GPU, not the TPM, is the root of hardware DRM(mjg59.dreamwidth.org)

399 points by DvdGiessen 1 year ago | 408 comments

Blog post by Matthew Garrett arguing that GPUs, not TPMs, are the real root of hardware DRM; the capture shows a Dreamwidth captcha page instead.

GPUs Go Brrr(hazyresearch.stanford.edu)

1056 points by nmstoker 2 years ago | 243 comments

Stanford Hazy Research blog post on what GPU hardware wants for efficient AI kernels, introducing the ThunderKittens CUDA kernel library.

We were wrong about GPUs(fly.io)

834 points by mxstbr 1 year ago | 529 comments

Fly.io blog post reflecting on how the company's bet on offering GPU machines to developers did not pan out as expected.

I replaced Windows with Linux and everything's going great(theverge.com)

647 points by rorylawless 9 months ago | 566 comments

A Verge first-person diary column by Nathan Edwards about replacing Windows with Linux on his gaming desktop, chapter one.

We bought the whole GPU, so we're damn well going to use the whole GPU(hazyresearch.stanford.edu)

472 points by sydriax 1 year ago | 94 comments

Stanford Hazy Research blog post introducing a throughput-optimized megakernel for tensor-parallel Llama-70B inference on H100 GPUs that overlaps compute, memory, and communication.

OpenFreeMap survived 100k requests per second(blog.hyperknot.com)

559 points by hyperknot 1 year ago | 120 comments

Blog post by Zsolt Ero describing how the OpenFreeMap tile service handled a sudden surge of 100,000 requests per second.