It’s popular to claim that AI (often meaning agentive LLMs) is now as bad as it will ever be.

I think it’s naive to underestimate how much worse anything designed by humans can become in the future.