Comment by ethbr1
11 hours ago
One of the worst things about talking with people who use lots of AI (less applicable to those who actually work on building AI) is their decreased resilience to opposing views.
AI has infinite time (and likely human evaluation incentive) to spend on couching pushback in the softest possible terms.
Humans outside of grade school honors classes generally don't have the time to preface "You're wrong" with "That's a brilliant thought, I see where you're going. How about we also consider an additional perspective..."
Seems to me that people with "decreased resilience to opposing views" have been increasing in number since well before LLMs.