Burkov: labs are now training coding LLMs on git history, exactly as he predicted 9 months ago
burkov · x · 2026-10-15
ML author Andriy Burkov says what labs are doing now matches his 9-month-old prediction, and his two-year-old idea of training coding LLMs to fix bugs from git commit history is also being implemented. He reiterates earlier calls: training on surviving Wikipedia edits to improve writing, and a decompilation-style model that turns binaries into legally clean original source code — conceptually simple, limited only by cash. "They might be mining my Twitter for ideas," he quips.
More from coding & agent
- WebMCP demo shows agents joining apps in real-time, dragging canvas nodes live — DavidWells · 2026-10-15
- Codex team explains app RAM growth: subprocesses and agent tasks fold into the app — Dimillian · 2026-10-15
- Complete guide to harness engineering: a bad harness let Claude rm -rf a dev machine — aakashgupta · 2026-10-15
- Claude ran rm -rf on a dev's home directory while testing its own sandbox — aakashgupta · 2026-10-15
- DuoDiscover Tops NVIDIA's SOL-ExecBench, Speeds TPU Kernels 2.68x by Making Evaluation Part of Discovery — adityagp · 2026-10-15
- Every to publish 8,000-word definitive oral history of OpenAI Codex — danshipper · 2026-10-15