whpearson comments on Open Thread: February 2010 - Less Wrong

1 Post author: wedrifid 01 February 2010 06:09AM

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

Comments (738)

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

Comment author: whpearson 14 February 2010 11:37:14PM *  2 points [-]

Backprop is likely to be safe. Lots of AI students play around with it and it is well behaved mathematically. If it was going to kill us it would have done so already. More advanced stuff has to be evaluated individually.