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

> "real" communism, one which is say, managed my an AI, with no one small group of humans in power, is probably going to be fair and equal.

Any rational agent, whether biological or artificial, A) possesses a set of preferences, and B) seeks to maximize its preferences. If an AI obtains its set of preferences for maximization from a group of privileged programmers or creators, then a small group of humans is still in power in your scenario. If the AI is responsible for deriving its own set of random preferences, then it is effectively operating as another individual in society with no deterministic guarantees on behavior, and to grant it governing authority is equivalent to unchecked dictatorship.

Our AI programs don't need to have their own preferences. Each rational agent can be assigned an assistant AI which will derive its preferences from the agent's already existent preferences.

It'll be like everyone has an AI genie, granting them wishes!

I believe by AI, he actually means oracle.

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