I agree with nouridin: Its probably the fear that AI would attack lethal entities that drive the AI tech-bros to slow down the development of better AU models.
I partly agree. I uses AI several times now, and to my experience its very fast and reasonable writing hundreds of DTOs/entities. They are withouit typos and created much faster then a human could it.
But if I ask him a question about architecture (for instance how to avoid race conditions, or how to implekmenet a role concept), the answers are a bit common sense. If I then tell him something like 'Wouldn't it be smarter when I do it this way?', it immediately agrees and does it 'this' way. So, my 5 cents: AI is good for routine work, but if it comes to intuition and experience, the human is still better.
Would be a nice thing, because atomic bombs wouldn't work at all. According to E=mc^2, the energy which would come out of a atomic fission would be hardly measurable.