Get Friedemann Zenke on Dwarkesh Podcast

AI is slamming into data, cost, and energy ceilings in 2026, and Friedemann Zenke is one of the few researchers showing a concrete path beyond scaling with brain-inspired learning. Just last month he headlined the Allen Institute’s Lake Conference with Latent Predictive Learning, a local rule that matches primate plasticity data and scales to deep and spiking networks.

In November 2025 he co-authored new work on “prospective neurons” that synchronize teaching signals across deep hierarchies, a practical fix for training systems with slow integrators that today’s backprop can’t handle. He is also shaping the spiking field’s agenda as part of the SNUFA 2025 program committee and with recent results on emergent assemblies in trained spiking networks.

Zenke has not appeared on the Dwarkesh Podcast before, so this is a rare chance to hear a leading spiking and plasticity theorist translate cutting-edge neuroscience into near-term, energy‑efficient AI. If you want smarter, cheaper, and more biologically grounded AI to happen sooner, support this nomination today.