We have two awesome new videos on MLX at #WWDC25 this year. - Learn all about MLX. - Learn all about running LLMs locally with MLX. Angelos Katharopoulos, Shashank Prasanna, myself, and others worked super hard to make these. Check them out and hope you find them useful!

you guys are SO creative man 😭😭😭😭😭😭😭😭😭😭😭😭😭

if you offer to make me eggs and i see the seasonings come out like you’re about to make wings… I’m cool lol you got it

i’m forming a model behavior research team ! i'm realizing llms are not alien artifacts, but rather craftsman works, like a braun record player or a compelling tolkien character join to do art & eval research on the internal frontier AND products that’ll reach billions

My mom told me she was at someone’s house and they invited her to stay for dinner. She said yes—until she saw the ingredients they pulled out for the ackee. Within 5 minutes, an urgent call came through…she had to bounce 💀

An important read 🚀

new decade, same verse

I'm hiring for Product roles in Gemini Google. Hit me up. DM's open!

We're hiring for the Claude Code team! In particular, we're looking for a Systems Engineer (job-boards.greenhouse.io/anthropic/jobs…) and an Eng Manager (job-boards.greenhouse.io/anthropic/jobs…). The Claude Code team is having a blast, come join us :)

🐐

Building a simulator for a complex system will expose design flaws and bugs you didn’t know you had.

You did miss the mark but ily anyway…this is hard. Good luck out here in these streets 🚀

Built different 🥴❤️

Hate when I tell ppl I’m Jamaican and they start acting stupid.

The only reason you'd offer a $1bn salary is if you believed you were incredibly close to cracking technology that will leave today's world in the dust. The only reason you'd turn down a $1bn salary is if you thought you might get there first. No one's following this enough.

"If you don't understand the code, your only recourse is to ask AI to fix it for you, which is like paying off credit card debt with another credit card."

You're absolutely right.

Qwen3-Coder-Flash runs quite fast on an M4 Max with mlx-lm. Running the 4-bit here, generated 4,467 tokens at >107 tokens/sec:

Our open models are here. Both of them. openai.com/open-models

Qwen3 be like