According to a Business Insider report, Tesla has made a major shift in how it trains its humanoid robot, Optimus. Instead of relying on motion-capture suits and VR headsets, the company is now moving to a vision-based approach, using video recordings of workers performing tasks.
This mirrors the methodology behind Tesla’s self-driving development, where massive video datasets are used to train neural networks to develop adaptive behavior. Workers now wear special helmets equipped with five Tesla-engineered cameras that capture detailed hand and finger movements from multiple angles, providing a rich dataset for training.
Leadership changes are also part of this transition — the program is now led by Ashok Elluswamy, Tesla’s Director of AI, following the departure of former project head Milan Kovac.
Experts highlight that vision-based learning could allow Optimus to generalize skills, meaning it can learn a wide range of tasks without needing individual training for each. However, they also caution that the lack of physical feedback (which teleoperation provides) may limit precision and adaptability in more complex scenarios.
Why does this matter? With this move, Tesla is accelerating the data collection process, aligning with Elon Musk’s philosophy that AI learns best through cameras — the same principle that underpins Tesla’s self-driving technology. The key question is whether the company can gather enough richly annotated video data for Optimus to master a wide spectrum of household and industrial tasks.
In brief: Tech World Highlights
- Google product lead Logan Kilpatrick posted a banana emoji on X, hinting that the “nano-banana” photo-editing model being tested on LM Arena likely comes from Google.
- OpenAI announced the launch of ChatGPT Go, a cheaper subscription specifically for India, priced at under $5 per month with payment available in local currency.
- ElevenLabs introduced Chat Mode, allowing users to build purely text-based conversational agents alongside its voice-focused systems.
- DeepSeek launched its V3.1 model with a larger context window, while Chinese media attributed the delay of the R2 release to CEO Liang Wenfeng’s “perfectionism.”
- Eight Sleep announced a new $100 million funding round, with plans to develop the world’s first “Sleep Agent” for proactive recovery and sleep optimization.
AI Trending Tools:
- GenFlow 2.0 – Baidu’s general-purpose AI agent for complex tasks.
- Stormy – An AI agent designed for influencer marketing.
- Imagen 4 – Google’s state-of-the-art image generation model, now generally available.
