'All your fancy new RL ideas were already tried in the 90s and 00s'
cephaloform · x · 2026-09-17
The author argues most 'new' RL ideas were already tried and written about in 90s/00s deep learning work, which has been swept aside because people only value results on chat-trained transformers — 'like throwing away bikes because we have planes.'
Related event: Transformer Hype Is Burying Decades of Deep Learning Research(2 posts)→
More from Research
- Protein Models ESMC and ESMFold2 Land on HuggingFace with NVIDIA-Built Fused Triton Kernels — AllThingsApx · 2026-09-17
- Panel Discussion on AI in Mathematical Research from Sept 15 Comes Highly Recommended — MindlessPapaya8463 · 2026-09-17
- OpenAI hints at millennium problem breakthrough ahead of Sept 29 DevDay — haider1 · 2026-09-17
- Vitalik: cybersecurity favors defense in the AI-hacking era, thanks to formal verification — jessi_cata · 2026-09-17
- PufferLib 5.0 hits 60M steps/sec single-GPU RL training, solves Breakout in under a second — jsuarez · 2026-09-17
- SGLang's Delta Router Replay slashes sync stalls in Kimi K2 agentic RL training — hsu_byron · 2026-09-17