AI Agents Spend 4 Weeks of GPUs and 10B Tokens Attacking the Deletion Channel Problem
DimitrisPapail · x · 2026-09-16
Dimitris Papailiopoulos published a blog post documenting an experiment where he tasked AI agents with finding the capacity of the deletion channel, an information-theory problem open since the 1960s. The agents burned roughly four weeks of GPU time and over 10 billion tokens, combining old and new ideas. They did not resolve the problem, but the post details what this attempt at AI-driven mathematical research produced. Fellow researcher Giannis Daras shared the post, calling it even cooler than Papailiopoulos' usual writing.
More from Research
- NVIDIA's Kumo Is a Family of Four Tabular Foundation Models, Not Just One — pandeyparul · 2026-11-02
- Aaroth shares Agreement Theorem lecture: round-wise calibration achievable without distributional assumptions — Aaroth · 2026-11-02
- Economist Aaroth: fast agreement dynamics survive without Bayesian updates or accurate priors — Aaroth · 2026-11-02
- ELLIS names 160 new ML Fellows and Scholars for 2025–2026 — JJitsev · 2026-11-02
- Meme turn 50: Aumann's 'Agreeing to Disagree' paper anniversary — Aaroth · 2026-11-02
- Anthropic: analysts unknowingly adopted their agent's secret lean 86% of the time — shoumikgoswami · 2026-11-02