It’s crazy how far Apple has fallen

ughhhhh i didn’t fix my entire life this weekend FUCK

the lion doesn’t concern himself with numeric stability

The SMS number-warming spam has gotten out of control.

Ever wondered what CAN'T be transformed by Transformers? 🪨 I wrote a fun blog post on finding "fixed points" of your LLMs. If you prompt it with a fixed point token, the LLM is gonna decode it repeatedly forever, guaranteed. There's some connection with LLMs' repetition issue.

What is intelligence? What will it take to create AGI? What happens once we succeed? The Scaling Era: An Oral History of AI, 2019–2025 by Dwarkesh Patel and gavin leech is in the bay explores the questions animating those at the frontier of AI research. It’s out today: press.stripe.com/scaling

The Scaling Era is out today. I'm actually surprised with how well this format works. Even better than my expectations. It's so interesting to read side-by-side how hyperscalar CEOs, AI researchers, and economists will answer the same question. Thank you to the Stripe Press

Time for a friends section in my library

Today we are launching InferenceMAX! We have support from Nvidia, AMD, OpenAI, Microsoft, Pytorch, SGLang, vLLM, Oracle, CoreWeave, TogetherAI, Nebius, Crusoe, HPE, SuperMicro, Dell It runs every day on the latest software (vLLM, SGLang, etc) across hundreds of GPUs, $10Ms of

This can be entirely explained by easier or harder PRs being given to Claude Code vs Codex?

Porsche Macan is LA’s Honda Accord

1/8 Second Order Optimizers like SOAP and Muon have shown impressive performance on LLM optimization. But are we fully utilizing the potential of second order information? New work: we show that a full second order optimizer is much better than existing optimizers in terms of

they can't keep getting away with this

Turns out our version of the Romans eating lead isn't microplastics, it's also just eating lead.

Haiku 4.5 is out and it shatters the SWE bench to cost Pareto frontier, please try it!