

AI is developing quickly. We want to understand not only what current systems can do, but where they are heading, what limits them, and what their development could mean for the systems around us.
Our research spans technical capabilities, the path toward more general intelligence, safety and alignment, and the broader consequences of increasingly capable AI.
Our research is early-stage. We will publish findings, open questions, and honest uncertainties as we go.
Question: What is intelligence?
Question: What should the image of AGI look like and its high-level shape.
These investigations change as we learn. We publish them when there is enough substance to share.