Hacker News new | past | comments | ask | show | jobs | submit login

What we really need is Local AI, not Cloud AI for such uses. It would suck to have to be internet-connected in order for your arm to work.

Want to take a trip into the wilderness? Lol. Power's out and the basement is flooding? Good luck, sucks to be you. I guess "lag" will become a valid excuse to being clumsy in real life.

Local vs connected is a no-brainer for this particular application.




Local inference with cloud learning isn’t an unreasonable model.


The only acceptable usage mode is a (preferably short) learning phase with the "cloud". After that you disconnect and preferably deactivate all wireless connectivity with a physical switch so that nobody can mess with it.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

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

Search: