Startup Elorian launches Glimpse, a visual thinking model that could automate robot data labeling
LinusEkenstam · x · 2026-10-17
- Startup Elorian has released its first "visual thinking" model Glimpse, which answers questions about images and video by reasoning through what it sees.
- Founder Andrew Dai says the model addresses years of frustration with models that ace single images but whiff on anything that moves or changes angle; Glimpse claims competitive frontier performance vs Google's Astra on key public visual reasoning benchmarks.
- Linus Ekenstam highlights the application angle: most teleoperation footage sits unlabelled because manual per-step segmentation is slow and expensive. If the model can watch clips and write the labels itself (what changed, what the robot did at each step), that backlog turns into training data — potentially transforming robotics data labeling.
More from Embodied
- Perplexity CEO builds interactive physics reconstruction of Tesla Roadster jump demo — AravSrinivas · 2026-10-17
- Professor recruiting one PhD student for Fall 2027 on agentic wearables — jasonwuishere · 2026-10-17
- Sandbar's Stream Ring nails whisper-quiet dictation across iOS and Mac in demo — thesephist · 2026-10-17
- U. Tokyo's TwinX robot: one wire links two wheeled-legged bots for morphing and teamwork — LexiLove · 2026-10-17
- Snapchat Specs pop-up in LA: tech is cool but the $2k glasses 'look ridiculous' — SimonBalmain · 2026-10-17
- IEEE profiles roboticist Peter Stone: robot soccer, service robots and being a generalist — PeterStone_TX · 2026-10-17