10% are working on AI capabilities, but only ~6 on pre-training at a frontier AI company (2 at Anthropic, 2 at Google DeepMind, 1 at Mistral AI, 1 extrapolated)
What about RL?
Why did you single out pre-training specifically?
The number I’d be interested in is the % that went on to work on capabilities at a frontier AI company.
What about RL?
Why did you single out pre-training specifically?
The number I’d be interested in is the % that went on to work on capabilities at a frontier AI company.
Sorry, I should have said “~6 on capabilities at a frontier AI company”.
What are some representative examples of the rest? I’m wondering if it’s:
AI wrappers like Cursor
Model training for entirely mundane stuff like image gen at Stablediffusion
Narrow AI like AlphaFold at Isomorphic
An AGI-ish project but not LLMs, e.g. a company that just made AlphaGo type stuff
General-purpose LLMs but not at a frontier lab (I would honestly count Mistral here)
Here are the AI capabilities organizations where MATS alumni are working (1 at each except for Anthropic and GDM, where there are 2 each):
Anthropic
Barcelona Supercomputing Cluster
Conduit Intelligence
Decart
EliseAI
Fractional AI
General Agents
Google DeepMind
iGent AI
Imbue
Integuide
Kayrros
Mecha Health
Mistral AI
MultiOn
Norm AI
NVIDIA
Palantir
Phonic
RunRL
Salesforce
Sandbar
Secondmind
Yantran
Alumni also work at these organizations, which might be classified as capabilities or safety-adjacent:
Freestyle Research
Leap Labs