Uncle Bob: It's Over • Hacker News Board

Re: Uncle Bob: It's Over

#91

Post by johnfn » Tue, May 05, 2026, 2:46 AM UTC

KolenCh wrote: post #83

johnfn wrote: post #34

Earlier quoted context omitted.

There are definitely tasks you can prompt an AI in 5 minutes that would take a whole day to do. One example is adding something to a CI pipeline and getting it to green (i.e. maybe you're adding your first ever e2e test), especially when your CI pipeline is painfully slow. e.g. if your pipeline takes 30 minutes to finish, and it takes around 10 tries to figure out all the random problems, that was easily a full day t…

Have you tracked your time and confirmed this? The reason I an ask is, it would felt like a 5 minutes task, but I track my time and found out often time I thought I’d just quickly check the progress made by the agents and it would easily becomes a 10, 15, or even a 30 minutes task.

The last time I did this I prompted the agent and went to sleep. When I woke up, CI was green. The agent had worked for 4 hours or something.

Re: Uncle Bob: It's Over

#92

Post by rstuart4133 » Tue, May 05, 2026, 5:04 AM UTC

He made a lot of assertions beyond the one in the headline: "Tests force better structured code" and "Mutation testing makes the LLM produce code with fewer bugs" are two examples. Most sounded plausible. But the delivery style of a monologue of rapid fire assertions made the entire thing very hard to accept.

A little bit of hard evidence would be nice. In fact it would be more than nice. Firm anchors in the form of published results of real experiments on how you can make LLMs produce different styles of code would be golden right now. I'm sure I'm not the only one flailing around looking for good ways to use these newfangled tools.

Re: Uncle Bob: It's Over

#93

Post by KolenCh » Wed, May 06, 2026, 5:47 PM UTC

johnfn wrote: post #91

KolenCh wrote: post #83

Earlier quoted context omitted.

Have you tracked your time and confirmed this? The reason I an ask is, it would felt like a 5 minutes task, but I track my time and found out often time I thought I’d just quickly check the progress made by the agents and it would easily becomes a 10, 15, or even a 30 minutes task.

The last time I did this I prompted the agent and went to sleep. When I woke up, CI was green. The agent had worked for 4 hours or something.

:)

I experienced both kind of sessions, one kind is like very complicated thing and it finish thousands of lines on its own for quite a while (long horizon problem), another kind is like a seemingly simple task that the agent done in a minute but then I need a few back and forth to get it right easily taking me 30 minutes of my time.

The former kind of experience can make us misjudge how much time we think a task would take us (with agents) to do. And then when the second kind happens, it would be quite disrupting as now we felt like it is delaying our progress.

So tracking the time taken when the second kind happens can help us calibrating what we can expect. I mean if we’re lucky it might take us no time but then we can’t expect being lucky all the time.