Hacker News new | past | comments | ask | show | jobs | submit login
AlphaZero taught itself to play three different games (arstechnica.com)
11 points by jonbaer 3 months ago | hide | past | web | favorite | 3 comments

Wonder why they still look at trivial problems, like playing limited games or doing analytical static research problems, but not solving current dynamic realworld problems, like e.g. driving cars, directing traffic control, ...

I’m curious... AlphaZero seems quite good at understanding closed systems, with deterministic rules.

People seem to think this will generalize to open systems, with no rules, like war and politics... why?

Probably becoz of too much science fiction.

Applications are open for YC Summer 2019

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