You're looking at Less Wrong's discussion board. This includes all posts, including those that haven't been promoted to the front page yet. For more information, see About Less Wrong.

paulfchristiano comments on [link] New essay summarizing some of my latest thoughts on AI safety - Less Wrong Discussion

14 Post author: Kaj_Sotala 01 November 2015 08:07AM

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

Article Navigation

Comments (27)
Tags: ai

Comments (27)

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

Comment author: paulfchristiano 15 November 2015 01:15:19AM 0 points [-]

delicate symmetry-breaking which can only come from either the training procedure or noise in the data, rather than the model itself

I'm still not convinced. The pointwise nonlinearities introduce a preferred basis, and cause the individual hidden units to be much more meaningful than linear combinations thereof.

Comment author: jsteinhardt 15 November 2015 07:48:24AM 0 points [-]

Yeah; I discussed this with some others and came to the same conclusion. I do still think that one should explain why the preferred basis ends up being as meaningful as it does, but agree that this is a much more minor objection.

Powered by Reddit Powered by Reddit