Post #1950030
2026-04-16 16:44 UTC
I am frustrated by the fact that Claude is actually decent at programming. I still think it’s not worth the costs wholly, because it’s still prone to mistakes and its code is awful.
My manager at work and I were talking about this and I mentioned that my attitude is that it can do about 80% of a given task. I basically feel like Claude and other AI LLM things can do, at best, around 80% of a given task.
My manager asked “What if 80% is enough?”
I had a response.
Replies (1)
-
@pandora_parrot@beach.city 2026-04-16 16:44
I said “Everyone can get that 80% from AI, including your competitors. What humans do in that 20% is what’s going to differentiate you and elevate you over your competition.” That 80% is the easiest part of the task, and it does it poorly/average at best. It’s not good at it. If you’re okay with mediocre, fine, but you’re going to get thwomped by someone that invests in humans doing human work much better than you can do it.