TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Learning human objectives by evaluating hypothetical behaviours

39 pointsby datashrimpover 5 years ago

1 comment

carapaceover 5 years ago
These systems (large, automated, computer-controlled, and capable of unsafe states) should be trained in such a way as to leverage each others&#x27; knowledge.<p>E.g., each <i>brand</i> of auto-auto (self-driving car) should share information with the others in re: safety and unexpected conditions. This should happen in (near-)real-time.