lateinteraction: frontier LLMs are getting eerily good at writing neural network weights
lateinteraction · x · 2026-11-03
lateinteraction responds to SchmidhuberAI's critique, noting his early work is cited and emphasizing the novel axis: how natively capable LLMs are at translating natural language into weights. The excerpted intro notes weights are traditionally produced by optimization, custom writing networks, or compilers from formal programs — and the current generation of frontier LLMs is starting to become eerily good at understanding and crafting neural network weights around natural-language goals.
More from Research
- VideoRebirth Report: Agent Harnesses for Agentic Video Generation, World Models and Recursive Self-Improvement — VideoRebirthAI · 2026-11-03
- Huawei's TinyAstral Distills Self-Evolving Harnesses into 4B Robot Agents, Wins Across Six Benchmarks — Kehong Gong · 2026-11-03
- KAIST's ReTrace Demystifies Computer-Use Agent Recovery, Finds Reflection and Rollback Fragile — kaist-ai · 2026-11-03
- Luxembourg study: empathetic framing pushes Gemini to human-level anxiety scores on GAD-7 — xiaohu · 2026-11-03
- Three wild predictions after MIT's weight-space programming work — Yuchenj_UW · 2026-11-03
- Mistral Large 4 tech report rejected by arXiv, team releases it on alphaXiv — _AndrewZhao · 2026-11-03