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

99% reliable is very convincing at looking like 100% reliable. People get comfortable with things pretty quickly.


Yes and every halving of that 1% error is more than twice as hard to do with neural net training. At some point more principled control algorithms are needed.




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

Search: