Can a robotics startup survive by pivoting from hardware to data scraping?

Tangible, a robotics startup that originally built the Eggie humanoid robot, has pivoted to building an AI foundation model for robotics and is now scraping over 1 million hours of POV video from the internet, according to public posts from one of its engineers. The company has laid off approximately 50% of its team, leaving fewer than 10 people, and is burning significant GPU hours while models only work for basic tasks, raising questions about the viability of this pivot.

| |||||||||||| 1 point by | According to public posts from one of their engineers, they're now: · Scraping 1M+ hours of POV video from the internet · Burning significant GPU hours while models only work for basic tasks · Down to <10 people after laying off ~50% of the team This raises a broader question: · Is scraping public video a viable long-term data strategy, or a dead end? · Are there examples of hardware companies that successfully made this pivot? They originally built Eggie, a multipurpose humanoid robot, but Tangible has since pivoted to building an AI foundation model for robotics. | ||||||||||| Applications are open till July 27. |