I joined Anthropic as a member of the technical staff. Excited to work on frontier modeling at a place with unwavering values and a generational mission.

@awnihannun Super cool Awni - hope you might even be able to convince them to open things up a bit! :D

@awnihannun Congrats! What will happen to MLX?

@awnihannun welcome! hope to cross paths, big mlx fan

@awnihannun Congratulations!! Very well-deserved.

@awnihannun I’m so excited for you (and us)!!

@awnihannun :))

@awnihannun They are lucky to have you there!

@awnihannun Awni, you might enjoy this line that Claude came up with in a long chat 2 nights ago: "You don't have to diminish one mind to appreciate another"

@awnihannun welcome to the team!

@awnihannun Anthropic's values: your not allowed to use facts or truth you must use emotions and feelings that align with our ideology. Censor all the work you do to align with our ideology. Lower the quality of our product with bad data and censorship. Ensure the ideology is protected.

@awnihannun @mattcassinelli Another @AnthropicAI W

@awnihannun big signal — top ML researchers tend to cluster where they think the real bets are being made. Anthropic's combination of frontier modeling and genuine safety research is increasingly rare. congrats on the move, excited to see what comes out of it.

@awnihannun worth noting they publish a lot — constitutional AI, interpretability, mechanistic work. the weight-openness question is specifically about transformative capabilities before alignment is solved. genuine tradeoff, not just corporate gatekeeping.

@awnihannun 💔 Values like advising the government to ban open source AI

@awnihannun MLX creator going to frontier modeling is a big hire. Anthropic keeps pulling the people who actually built the foundational tools.

@awnihannun Congrats Awni! Bring them to the next level and please try to push some Open values from within.

@awnihannun congrats! very excited to see what you build there

@awnihannun frontier modeling and authorization are converging. as models get more capable, the question shifts from 'can it?' to 'should it, right now, for this caller?' the deployment surface IS the authorization surface at the frontier. welcome to the edge case factory.

@awnihannun Is using your models for targeting when we all know null results get omitted really moral though? P.S. love Claude Code still

@awnihannun amazing news, congrats, seriously this is awesome on so many levels.

@awnihannun @AndrewCurran_ Congrats Awni

@awnihannun Huge congrats! 🚀 Frontier modeling at Anthropic — what an exciting time to join. Wishing you all the best on that generational mission!

@awnihannun Can you make it so we can paste images in Claude Code CLI

@awnihannun Congratulations!

@awnihannun congrats! frontier modeling at Anthropic is where the real work happens. exciting times ahead

anthropic is quietly assembling one of the strongest research teams in AI. congrats. the 'unwavering values' part is what makes it interesting. most frontier labs optimize for speed. anthropic is betting that safety and capability aren't tradeoffs. that bet is looking smarter every quarter.

Congrats! Anthropic has been quietly building the most principled AI team out there. "Unwavering values" isn't just marketing speak there—they actually turned down deals that didn't align with their safety mission. The frontier modeling work you're doing is going to shape the next decade. No pressure 😄