AI PhD: CS academia is dead — ICLR submissions are just swapped-in base models
ZeeshanZiaML · x · 2026-09-26
An AI PhD says CS academia is 'dead': during ICLR reviewer bidding he found papers recycling the same recipe he learned two years ago — swap in a newer video gen base model and ship. He suspects many of the 60k submissions were AI-conceived, and argues industry (World Labs' Atlas on 4D scenes, GPT-6 Astra demos) has left academia far behind. Meta's Dimitris Papail quotes it: basic research is in a rough spot but vital — go after problems so hard they'd have seemed hubristic a year ago.
Related event: AI PhD Slams ICLR Submissions as Recycled, Says Academia Lags Industry(2 posts)→
More from AGI Musings
- User publicly pleads against deprecating Opus 5, citing projects built together — repligate · 2026-09-26
- 'Remember Web 3.0?' Industry vet says much of today's AI hype will share its fate — JFPuget · 2026-09-26
- Investor argues agents should buy answers, not web fragments, challenging Cloudflare's microtransaction model — Ronangmi · 2026-09-26
- Recursive Self-Improvement Is Misunderstood: Self-Improvement Alone Is Just Iteration — emax · 2026-09-26
- Jensen Huang dismisses scaling laws, calls AGI "just marketing" in Hard Fork interview — teortaxesTex · 2026-09-26
- By Ricardo's law of comparative advantage, should humans now do busywork for AI agents? — arjunrajlab · 2026-09-26