Gzip, RE-PAIR Grammar Induction and the Case That Deep Learning Is Just the Cerebellum
ryunuck · x · 2026-09-21
The author argues classical compression and deep learning should be mixed to cover each other's failings, claiming we only have "the cerebellum, but not the cerebral cortex." Points include: gzip is a better lossless compressor where an RL equivalent could be invented, and grammar induction algorithms like RE-PAIR over a byte tape remain unexplored compression avenues.
More from Research
- Founder says frontier lab's Jev copies his 2025 non-autoregressive decision model paper and open weights — abhijithneil · 2026-09-21
- Schmidhuber: LeCun's JEPA is essentially his 1992 Predictability Maximization system — SchmidhuberAI · 2026-09-21
- Score Centering is secretly a STE: new fix targets LLM RL training instability — brandondamos · 2026-09-21
- Most Researchers Still Treat Transformers as Black Boxes, and Public Understanding Is Decades Away — gerardsans · 2026-09-21
- Paper pinpoints EOS token mismatch as the root failure mode of on-policy distillation — tw_killian · 2026-09-21
- AI Haiku Study: GPT-5 and Gemini 2.5 Poems Indistinguishable From Human Work — s_scardapane · 2026-09-21