Comment by baroudi
8 hours ago
> But there are serious limits. [Your coding agent] will lie to you, they don't really understand things, and they often generate bad code.
As for lies and bad code, it didn't appear with AI. Humans lied and produced bad code before AI.
How does the author empirically know AI does not understand? And if it does not understand right now, is a machine fundamentally unable to understand? Is understanding an exclusive human ability? Is it because machines lack a soul? It sounds quite dualistic (Descartes'view that mind and body and fundamentally different).
Don't get me know, I think right now, AI is less a good at understanding humans than other humans (or even dogs) in many contexts because it has no access to non verbal signals. But in the context of building software, it is good enough and I don't see why a machine should not be able to understand humans.
I have had an interesting experience just recently.
I hired back on at a company I used to work at and found they had contracted work to another former employee who was handed the code from a rest api I had written and a web app that used it. The task was to write an android app that interacted with the api.
He ran it through an agentic coding assistant and got out api scaffolding and basic UI.
Looking it over, I couldn't shake the feeling I was looking at my code, just ported to kotlin. I was seeing my idiosyncrasies everywhere. It was kind of surreal.
I was familiar with the dev's work who did it and it was nothing like his prior work, but it's been years since I have seen him, so who knows.
I couldn't help but admit it was a good foundation to start building on.
I told the pm they were likely overpaying significantly for an agentic coding assistant and only getting access to it for a few hours a month. This same organization recoiled in terror when I pointed out the cost of a claude code subscription once.