Get Nikolaus Kriegeskorte on Dwarkesh Podcast

Nikolaus Kriegeskorte matters right now because his lab just won a new NIH BRAIN Initiative grant on November 5, 2025 to advance representational geometry methods that directly compare AI models with neural activity, the kind of work that can immediately sharpen how we evaluate and steer frontier models. In October 2025 he also coauthored a paper introducing an encoder decoder transformer that uses generative models to map category selectivity across the whole brain, a concrete step toward brain informed interpretability at exactly the moment society is deciding how much to trust powerful systems.

Dwarkesh’s show is where cutting edge AI ideas get road tested for the people building them, and he has the ear of Silicon Valley according to The Economist, so Kriegeskorte explaining brain to model alignment there could influence research priorities in 2026. As far as public records show, this would be his first appearance on the Dwarkesh Podcast as of January 20, 2026.

If you care about making AI safer and smarter using the brain’s playbook, back this now so the conversation happens in the most consequential room in tech. Please support this nomination today.