As others have said in other posts, the cat's out of the bag. No matter if/when/how bad the AI crash is, if computing is still viable after then there will be AI.

I think eventually it will emerge that the highest quality AI code will be produced by already experienced developers, who can identify when it's produced slop and fix it, and that will be the case no matter how good the models get.

I of think of it somewhat like managing a team of interns. You give them instructions and they give you code. You still have to know enough to identify if that code meets the requirements and is of good quality.