MPI's Michael Black: Learning About VIGA at ECCV 2026 Means You're Late to the Party
Michael_J_Black · x · 2026-09-09
Michael Black of Max Planck Institute reflects on the role of academic computer vision research as large models grow more powerful, ahead of ECCV 2026 — a conference he's attended since 1992. His paper VIGA (Vision-as-Inverse-Graphics Agent, which turns an image into a 3D Blender scene) changed the field's thinking on inverse graphics, but he notes that impact happened entirely via arXiv and social media, with follow-on work now driven outside academia by rapidly evolving large-model capabilities. "If you are learning about VIGA at ECCV, you are getting to the party late," he writes, while inviting discussion at his poster and raising questions like whether GPT-6 Astra represents a step change and how academics can still have impact today.
More from AGI Musings
- "Humans can't keep control" — the contrarian optimism argument around ASI — flowersslop · 2026-09-18
- AI circle's sharpest quip yet: you can't talk a man off doom if his ego depends on it — basedjensen · 2026-09-18
- "This ship can't sink": mocking ASI containment optimism with the Titanic meme — flowersslop · 2026-09-18
- Jagged RSI: Recursive self-improvement will keep generalizing as FLOPs come online — Justin_Halford_ · 2026-09-18
- OpenAI's math: ~$15B from users, trillions from investors betting on automating all labor — ComfortableSpeech302 · 2026-09-18
- Public AI awareness jumped straight from Ghibli images to existential doom — dbreunig · 2026-09-18