←back to thread

AI should only run as fast as we can catch up

(higashi.blog)
196 points yuedongze | 2 comments | 08 Dec 25 17:38 UTC | HN request time: 0.42s | source
Show context
gaigalas ◴[09 Dec 25 08:06 UTC] No.46202475[source]▶
Prompt engineering: just basic articulation skills.

Context engineering: just basic organization skills.

Verification engineering: just basic quality assurance skills.

And so on...

---

"Eric" will never be able to fully use AI for development because he lacks knowledge about even the most basic aspects of the developer's job. He's a PM after all.

I understand that the idea of turning everyone into instant developers is super attractive. However, you can't cheat learning. If you give an edge to non-developers for development tasks, it means you will give an even sharper edge to actual developers.

replies(1): >>46202496 #
1. booleandilemma ◴[09 Dec 25 08:09 UTC] No.46202496[source]▶
This is true. I've been anti-ai but I started using it recently as an alternative to stack overflow (because google is shoving it down my mouth via search results). It's pretty effective. It does get things wrong from time to time, but then I just fix it up manually. I can't claim it's making me 100x more productive or anything like that. It's just a nice alternative to scrolling through SO answers and looking for the one with the green checkmark.

I still find it sad when people use it for prose though.

replies(1): >>46202576 #
2. gaigalas ◴[09 Dec 25 08:25 UTC] No.46202576[source]▶
If an agent gets things wrong you should stop it and correct it instead.

Sometimes the correction will cost more than starting from scratch. In those cases, you start from scratch.

You do things manually only when novel work is required (the model is unlikely to be trained with the knowledge). The more novel the thing you're doing, the more manual things you have to do.

Identifying "cost of refactoring", and "is this novel?" are also developer skills, so, no formula here. You have to know.