gwern comments on Open Thread, May 1-15, 2012 - Less Wrong

7 Post author: OpenThreadGuy 01 May 2012 04:14AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (264)
Tags: open thread open_thread

Comments (264)

You are viewing a single comment's thread. Show more comments above.

Comment author: gwern 01 May 2012 02:51:20PM 5 points [-]

First is this evidence that randomized rule systems that are complicated enough to be interesting are also likely to allow some sort of drastic increase in effective abilities using some sort of loopholes? (essentially going FOOM in a general sense).

Can you analogize this to being Turing-complete? One thing esoteric languages - and security research! - teaches is that the damndest things can be Turing-complete. (For example, return-into-libc attacks or Wang tiles.)

Comment author: David_Gerard 01 May 2012 03:36:08PM *  1 point [-]

the damndest things can be Turing-complete

Yep. Which is why letting a domain-specific language reach Turing-completeness is a danger, because when you can do something you will soon have to do it. I've ranted on this before.

Idle speculation: I wonder if this is analogous to the intelligence increase from chimps to humans. Not Turing-completeness precisely, but some similar opening into a new world of possibility, an open door to a previously unreachable area of conceptspace.

Powered by Reddit Powered by Reddit