New whitepaper: The Discovery Engine

We’ve just published a whitepaper that outlines the core ideas behind Leap’s approach to autonomous scientific discovery. It brings together a lot of thinking I’ve been doing over the past year – on what science should look like in the age of AI, what stands in the way of massively accelerated breakthroughs, and why LLMs don't seem to be able to help that much.

It's not a technical deep dive or a product spec, though it does outline what we've built so far. It’s more about what motivates us to build this thing specifically, when everyone else is leaning on language models – and why I think this is the best and only way to access the full potential of AI for science.

You can read it here: The Discovery Engine.