I just listened to a conference talk on AI. I suppose it was meant to be positive about agentic coding.
But the talk itself said:
– AI is used a lot, but developers trust it little
– Trust in AI coding output is falling, not rising
– AI generated code is generally insecure and poor standard.
Does this not… like… tell us something about these tools?
Do people see these stats and think “Oh yes – that sounds like something we should use”?