The recipe to solve any perception task: a few hundred examples and a fine-tuned VLM
giffmana · x · 2026-11-15
A shared recipe for solving "any perception task": pick a task, collect a couple hundred examples in an afternoon, fine-tune a large pre-trained VLM, RL-tune only if needed, then deploy. giffmana highlights point 2 — how little data it takes — as the crucial insight.
More from coding & agent
- Indie dev shows off T3 Code dashboard with 452 settled agent threads — yihui_indie · 2026-11-15
- Amazon planning World Product Graph to index the web's products for AI shopping agents — jomic01 · 2026-11-15
- Google Labs open-sources 15 AI Design Skills that turn prompts into React UI, 9k GitHub stars — tokumin · 2026-11-15
- 20 GitHub repos turn Fable 5.2 into a full motion studio — 3D, FX, sound, export — PrajwalTomar_ · 2026-11-15
- 79k-Star AI Engineering Course Ships Its Own Agent Tutor for Claude Code and Cursor — ghumare64 · 2026-11-15
- Flappy Bird built with Hala Play + Claude: all art and audio AI-generated — aziz4ai · 2026-11-15