Industry's embodied CV data dwarfs academia's, chart shows log-scale gap
ducha_aiki · x · 2026-09-09
Sharing a chart from Jakob Engel (log-scale Y axis) showing the massive gap in embodied CV data volume available to academia versus industry — a disparity that helps explain why end-to-end robot learning is increasingly hard to replicate outside large industrial labs.
More from Embodied
- One model that codes, uses computers and controls robots: a real step toward AGI — k7agar · 2026-09-10
- GRID launches Auto-Engineering: agents build robot skills, first skill in ~4 hours — akapoor_av8r · 2026-09-10
- GPT-6 shows 'incredible' physical-space instruction following on real robots, early tests show — k7agar · 2026-09-10
- GPT-6 Astra Shows Impressive Instruction Following on Robots — k7agar · 2026-09-10
- Binary bits: a $1M robot that breaks every few days doesn't count as 'capable' — binarybits · 2026-09-10
- Tesla Cybercabs spotted at Giga Texas test track alongside Model Y L — yunta_tsai · 2026-09-10