DeepMind's Koray: no single test will declare AGI, progress means less spiky models

OfficialLoganK · x · 2026-11-20

Google DeepMind SVP Koray Kavukcuoglu, in a conversation with Hannah Fry, argues there won't be a single test or benchmark that tells us AGI has arrived — progress looks like models becoming less spiky and more reliable across daily life.

The podcast also covers DeepMind's evolution, products as a path to AGI, and architectures. Logan Kilpatrick shared it calling it a great conversation.

Related event: DeepMind's Kavukcuoglu: No Single Benchmark Will Define AGI(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →