Mapping the Future of Science with AI | How Jonah Lynch Uses LLMs to Discover What Doesn’t Exist Yet
@NS Library
What if you could predict where science is going — before it gets there?
Jonah Lynch is a polymath: physicist, educator, AI researcher, and founder of Innovation Lens — a platform that uses LLMs and latent space modeling to uncover emerging scientific topics before they even have a name.
He holds a PhD in AI applied to historiography, has lectured at Oxford, MIT, and the University of Chicago, and has built tools that map gods in Mesopotamian texts, predict hospital overflows, and translate abstract paintings into music using neural networks.
In this live session, Jonah will lead a compact 15–20 minute deep dive and demo of innovationlens.org, followed by open conversation.
He’ll cover:
- Information overload as a bottleneck for science
- LLMs as knowledge maps
- What latent space reveals about discovery
- How to generate novel, actionable research hypotheses
This talk is part of the fast.ai Learnathon — a 3-day async sprint to build your own AI model — but is open to everyone.