= 27d567bc2d48d9f40e9ccc20ea370b15)
Vladimir_Nesov comments on Open Thread: March 2010, part 2 - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (334)
One career path I'm sort of musing about is working to create military robots. After all, the goals in designing a military robot are similar to those in designing Friendly AI: the robot must know somehow who it's okay to harm and what "harm" is.
Does this seem like a good sort of career path for someone interested in Friendly AI?
No. FAI is about figuring out how to implement precise preference, not an approximation of it appropriate for non-magical environments. Requires completely different tools.
It seems that to work on FAI, one has to become mathematician and theoretical computer scientist (whatever the actual career).
What do you mean by "non-magical environments"?
I gave a link! A non-magical environment gives limited expressive power, so there are few surprising situations that given heuristics don't capture. With enough testing and debugging, you may get your weakly intelligent robot to behave. Where more possibilities are open, you have to get preference exactly, or the decisions will be obviously wrong (see The Hidden Complexity of Wishes).
Your terminology was unclear but this definition is not - I would tend to call it an "organic" environment.