All stories
5 stories found
AI Systems Are on Pace for a Perfect Score at the 2026 International Math Olympiad
Prediction markets now put a perfect AI score at the 2026 International Math Olympiad above 90%, a milestone that would reset how researchers talk about frontier model reasoning.
A New AI Model 'Thinks in Images, Not Just Words'
A new AI model uses images as part of its reasoning process instead of translating everything back to text, hinting at where multimodal frontier research is going next.
Demis Hassabis on DeepMind's Comeback: From 'Behind the Curve' to a Nobel Prize
In a rare long-form interview, Demis Hassabis walks through DeepMind's arc from underdog to Nobel laureate — and explains what he thinks the AI industry keeps getting wrong.
Anthropic's Claude 'J-Lens Consciousness' Debate, Debunked
A close-read piece pushes back on the 'Claude is conscious' framing, arguing the underlying model-welfare research is nuanced and being flattened by headlines.
Epoch AI's Latest Data on Frontier Capabilities: The Frontier Is Getting Denser, Not Just Higher
Epoch AI's latest capability data set shows the frontier is getting denser, not just higher — with real implications for how labs pick benchmarks and how buyers pick models.