The Mathematics of Training LLMs(latent.space)

171 points by FanaHOVA 3 years ago | 58 comments

The AI Content Flippening(latent.space)

31 points by swyx 3 years ago | 36 comments

Replit's new Code LLM: Open Source, 77% smaller than Codex, trained in 1 week(latent.space)

848 points by swyx 3 years ago | 214 comments

Latent Space podcast episode with Replit's Reza Shabani about training Replit's open-source code LLM in one week and evaluating it.

Petaflops to the People: From Personal Compute Cluster to Person of Compute(latent.space)

47 points by tosh 3 years ago | 7 comments

RAG is a hack(latent.space)

36 points by FanaHOVA 3 years ago | 16 comments

$2 H100s: How the GPU Rental Bubble Burst(latent.space)

325 points by swyx 2 years ago | 210 comments

Latent Space guest essay by Eugene Cheah on how H100 GPU rental prices fell from about $8/hr to under $2/hr and what caused the bubble to burst.