OurWord.
1 read 中文

The world is too loud. Read what matters.

Google DeepMind

Official DeepMind; slow to update but all firsthand research

Deep reads here1 / 50 episodes
Cadence~1 every 21.8 days
Latest2026-08-26
TopicsAI & Tech
PriorityT2
44:40
Google DeepMind: The Podcast 0826

AI's next step isn't bigger models, it's learning to admit uncertainty

DeepMind researcher Zubin draws on 30 years of research to show that today's large models make things up confidently because they lack any explicit representation of probability; Bayesian updating, forecast ensembles and calibration are what reliable AI and AGI actually turn on.

7 Points 5 Quotes UncertaintyBayesian methods