I am excited and beyond grateful to be joining the founding team of Thinking Machines Lab, a new AI research and product company led by ๐ @miramurati , alongside a remarkable group of colleagues from @OpenAI , @AnthropicAI , https://t.co/JQANJnW7BP, @GoogleDeepMind, @MistralAI, and @Meta.
We are building:
๐ท Tools that let anyone adapt AI systems to work for their specific needs
๐ท Foundations for more capable AI
๐ท Open science that helps the whole field understand and improve these systems
You can read more about @thinkymachines below. Please consider joining us!
https://t.co/XMsoVlzeup
We built the lab that's able to go from AI-led drug design to data in 24h.
GPT-8 won't be bottlenecked by intelligence. It needs a biological compute layer.
This is Capable. We're turning AI capabilities into human capabilities--starting with short-sleeper peptides.
Releasing weights indiscriminately isn't safe. Neither is keeping capable models inside a few labs.
We think there's a path between them. We haven't mapped all of it. Our new post covers the part we can see: how we assessed Inkling, and why access should widen in stages.
https://t.co/zGFbuXNr0Y
We share our approach to open-weights releases, how we assessed Inkling, why safety depends on both the model and the ecosystem it enters, and how testing, staged access, and stronger defenses can create a path toward greater openness.
Whereas I felt like it took a village to release inkling, inkling-small felt much more routine ๐ We just took the pipeline used for Inkling, passed in a smaller model, and voila - new model! Inkling small benefited quite a bit vs Inkling from some minor improvements, but there's still so much more left in the tank...
Inkling-Small is comparable to Inkling at a quarter the size. Weights are open, fine-tunable on Tinker today. Look forward to seeing what people make with it.
last night we had the privilege of hearing @miramurati and @mkratsios47 share their distinct approaches to the belief that breakthrough innovation comes from empowering individual experts
The knowledge that makes AI useful is diffused. It lives with scientists, engineers, clinicians, firms. For AI to benefit from distributed knowledge, it must itself be distributed. Agree with Jensen that this is a future worth building.
Inkling from @thinkymachines on ARC-AGI (Verified)
- ARC-AGI-2: 36.5%, $0.64/task
- ARC-AGI-1: 79.5%, $0.30/task
As of today, Inkling is the highest-scoring open-weight model evaluated by ARC Prize on both ARC-AGI-1 and ARC-AGI-2.
After years of dumb posts with bad economics and 0 knowledge about the world of work (what they call "computer use") from AI lab bosses, the Thinking Machines manifesto is refreshing in its acknowledgements of what the collective work of humans in markets is about: the use of decentralized knowledge. Beautiful project. A lab with a vision to root for.
https://t.co/Et3dRNVj7n
Built a podcast clipping app with Inkling from @thinkymachines โจ
The model is exceptional at reasoning over long-form audio - so I have it listen to full episodes and direct FFMPEG on which clips to cut.
You can have it choose the best moments or search for specific topics ๐