I think the position has shifted to "let's pretend this problem doesn't exist because the AI market is too big to fail"