Maybe they have seen AI fail for so long due to poor performance that it makes AI safety experts seem like they are vastly overestimating what AI can do?

Source: [Hacker News](https://news.ycombinator.com/item?id=49111613)

Sponsored