Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

We are building the products and services that bring AI into the physical world.

Similar to living organisms, and unlike traditional AI, which operate in abstract digital environments, embodied AI enables robots to learn by doing.

It seems very plausible to me that the next step function change in AI is if you can connect perception, cognition, and action through direct physical interaction with the world.

But more than anything, what stands out to me is the optimism in improving life for humans as articulated by their guiding principles.



Tesla could possibly claim to participate in this shiny future, if only they had the geospatial data that Google has. Lack of real world data is why FSD sucks.


Lack of real world data is why FSD sucks.

All the data in the world can't really fix FSD because driving is a dynamic activity filled with unique, constantly evolving conditions.

FSD sucks because it is poorly engineered. It has no redundancy. It is built around a single point of failure --- vision. If this fails for any reason (sunlight, darkness, rain); it is essentially driving blind.

Tesla's engineers understand this but apparently Muck doesn't --- his main skill is marketing and salesmanship (targeting investors) --- not engineering.


All true. I am pointing out another dimension in which FSD faces a large obstacle to working well enough to be a real product.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: