00:00I think you also see a real fear
00:05that some of these companies
00:07or some companies developing AI
00:08could get too much power
00:09and be able to sort of exert undue influence
00:13on the economy, push a worldview out on people.
00:16And I think the world is right to be afraid of this.
00:19So we have these two big challenges.
00:21We have the potential of a loss of control accident
00:24or some other serious thing that could go wrong.
00:26We have the potential of way too much power concentration
00:29and people fear that those developing AI
00:32could exert their worldview.
00:35And there is a sort of narrow path
00:36we have to navigate with pragmatism,
00:41with steadiness,
00:44with the ability for the world
00:46to trust that we will make dependable decisions,
00:49consistent decisions through this.
00:51And I think people are saying,
00:53well, this now seems pretty important.
00:55Like, let's get it right.
00:57These models have gotten so good so fast
00:59that we need to treat the alignment
01:02and monitoring and security
01:03with a new level of rigor.
01:05And we have to be willing
01:06to pace our development of capabilities
01:09such that capabilities always...
01:13such that alignment, safety, and monitoring
01:15are always ahead of capabilities.
01:16in place.
01:17Thank you.
01:18Thank you.
01:19Thank you.
01:19Gracias por ver el video
Comentarios