grief is a natural response to the possibility of loss. it doesn't mean you're giving up, or that there's no hope. in fact on the other side of grief is often more clarity, resilience, and determination to protect what is good.
working extremely hard while avoiding our feelings makes us brittle, and we'll snap under the pressure of the coming years. by default, I expect the rate of burnout among AI safety / alignment / etc researchers to skyrocket in the next 2-3 years. this also produces a kind of "fight or flight" response at the group level, and I'm skeptical of any "solutions" that we come up with from this place.
a mood I'm really missing in the current AI discourse is grief
yes things might go terribly and yes we might see glories beyond imagining but no matter what, we will lose much of what it has meant to be human, forever. I'd like to be with that grief more, and held in it.
for those at IASEAI in Paris, I'm co-hosting a couple of post-IASEAI decompression events on Friday (an interactive experiment in collective intelligence, and a private concert). DM me for details!
totally surreal to see the intensity of reactions to this post. I could be wrong, but I can't remember any other AI lab departure letters being this scrutinized (even from much more famous researchers).
particularly hilarious to see people say Mrinank "must be a nightmare to work with" or some such thing, who have clearly never met him. Mrinank is one of the most wholesome and genuine beings I've ever interacted with. there are few people I'd rather work with in any setting.
I guess the vagueness of the post makes it a perfect screen for people to project their own ideas, beliefs, and insecurities on. a lot of the hateful reactions seem to be from "builder" types, as far as I can tell. not sure what that means exactly.
it's been a couple of years since my AI lab departure, and I decided not to post anything at the time, but if I had it would have been totally inconceivable to me that it would get this much attention. I guess a lot has happened since 2024. now that I've seen the reaction, it's kind of obvious (there haven't been as many Anthropic departures, plus it's unclear if he's trying to signal something really important). but still, it's kind of a bubble popping moment for me
Sculptor: the missing UI for Claude Code 🎨
Imagine running 5 Claudes in parallel, safely in containers, while you stay in flow. Then bring their work straight into your IDE to test/edit together.
This is how one developer ships like a team. Try it with Sonnet 4.5!
@RichDecibels Yes, I am doing a lot of work around it in collective trauma processes with Thomas Hübl. Remembering is intense and beautiful, heartbreaking and deeply connecting at the same time. When generations later Jews and Germans among others come together to feel what has not been felt
Hello world 😀
Thrilled to share the first episode of my podcast, Evolving with AI! Had a touching conversation with @BayoAkomolafe on GenAI, befriending the monstrous, and embracing awkwardness. Hope you find it inspiring.
open.spotify.com/show/0VyMYIFkD…
4K Followers 4K FollowingTrisolarian Panspecies Anachronist / Semantic Ghost Hunter
dms are welcome but will be only be answered during stable eras when rehydrated
11K Followers 1K Followingplaying & learning with nascent intelligences, particularly my baby daughter & LLMs. devout whatever-I-am, praying for more clarity of what godshape to pray to
134K Followers 1K FollowingSemiAnalysis
Boutique AI Infrastructure Research and Consulting
DMs are open for consulting, quotes, or to talk shop,
Opinions my own
23K Followers 316 FollowingDoing RL @AnthropicAI. Formerly VP of Research, Head of Post-Training @OpenAI. PhD with Aaron Courville and Marc Bellmare at Mila.
106K Followers 426 Followingprofessor of computer science @Stanford @stanfordnlp, co-founder of @togethercompute, creator of https://t.co/7R5THVogW2, co-founder of @simile_ai, pianist
257K Followers 10 FollowingWe’ll help you make it like nobody’s business. Multimodal media generation and editing tools to get your idea to production. Self-deploy? 👍 Need a partner? 🤝
256K Followers 178 FollowingCo-founder of Thinking Machines Lab @thinkymachines; Ex-VP, AI Safety & robotics, applied research @OpenAI; Author of Lil'Log
534K Followers 2K FollowingPolyagentmorous ClawFather. Came back from retirement to mess with AI and help a lobster take over the world.
@OpenClaw🦞 + @OpenAI
876K Followers 692 FollowingCo-Founder, LinkedIn. Investor. MSFT Board Member. Building an LLM to discover cures for cancer: @manas_co. Most importantly: Proud American.
10K Followers 1K FollowingAssistant Professor at NUS. Scaling cooperation for an increasingly automated future. PhD @ MIT ProbComp / CoCoSci. Pronouns: 祂/伊
7K Followers 476 Followingwith gentle fierceness | metabolising god | https://t.co/ITg08u2uvP | avoid twitter DMing me if time-sensitive; I do not check often
273 Followers 386 FollowingPost-doctoral Researcher at BIFOLD / TU Berlin interested in interpretability and analysis of language models. Guest researcher at DFKI Berlin.
22K Followers 5K FollowingRaising AI risk awareness at https://t.co/Fat9r8oGp0
AI prof at Mila.
Formerly Cambridge, DeepMind, UK AISI.
https://t.co/KKWOE9xKZB