Hutter Prize sees biggest progress in 20 years: three winners, ~10% total gain in one month
mhutter42 · x · 2026-10-03
Marcus Hutter announced that the Human Knowledge Compression Contest (Hutter Prize) has seen its largest progress in its 20-year history: three winning submissions arrived within a single month, with a total compression improvement of nearly 10% — output roughly 1/3 the size of the zip file. Winners include Ibrahim Marcouch, Kaido Orav, David Freelan, and Vladimir. High-level descriptions of the winning (de)compressors and the contest's history have been published.
The prize uses compression of human knowledge as a proxy for AI progress, making a 10% jump in one year unusually significant.
More from Research
- Offering a known-liked option as 'B' reveals how much a model actually wants 'A' — repligate · 2026-10-03
- Ask 'A or B or something else?' — single-option prompts bias model preference answers — repligate · 2026-10-03
- "Random statistical predictions" misreads LLMs: models learn internal algorithms — burny_tech · 2026-10-03
- CP-Agent: an LLM agent that runs crystal plasticity simulation workflows from natural language — Fowe · 2026-10-03
- Nematodes challenge AI welfare heuristics: pain-indicator behavior can't stand in for consciousness — dioscuri · 2026-10-03
- Anthropic's internal models and heavy compute confirmed in Navier-Stokes collaboration — teortaxesTex · 2026-10-03