Less human AI agents, please • Hacker News Board

Re: Less human AI agents, please

#181

Post by K0balt » Thu, Apr 23, 2026, 1:11 PM UTC

pythonaut_16 wrote: post #94

felipeerias wrote: post #51

Earlier quoted context omitted.

Claude 4.7 broke something while we were working on several failing tests and justified itself like this: > That's a behavior narrowing I introduced for simplicity. It isn't covered by the failing tests, so you wouldn't have noticed — but strictly speaking, [functionality] was working before and now isn't. I know that a LLM can not understand its own internal state nor explain its own decisions accurately. And yet, I…

> strictly speaking, it was working before and now it isn't I've been seeing more things like this lately. It's doing the weird kind of passive deflection that's very funny when in the abstract and very frustrating when it happens to you.

The thing to remember is that LLMs deeply model human behavior. If you want them to do their best work, you need to treat them like a collaborator and get them”invested” in the work and the outcome. I use an onboarding process with every new context and maintain an environment where a human would likely feel invested in the work and the outcomes. For me, it prevents a host of failure modes, and code quality has markedly improved.

Re: Less human AI agents, please

#182

Post by standardly » Thu, Apr 23, 2026, 6:28 PM UTC

CPLX wrote: post #78

gregates wrote: post #15

The version of this I encounter literally every day is: I ask my coding agent to do some tedious, extremely well-specified refactor, such as (to give a concrete real life example) changing a commonly used fn to take a locale parameter, because it will soon need to be locale-aware. I am very clear — we are not actually changing any behavior, just the fn signature. In fact, at all call sites, I want it to specify a def…

I've have a different version of the same thing. My pet peeve is that it constantly interprets questions as instructions. For example, it does a bunch of stuff, and I look at it and I say, "Did we already decide to do [different approach]" And then it runs around and says, "Oh yeah," and then it does a thousand more steps and undoes does what it just did and gets itself into a tangle. Meanwhile, I asked it a question…

Solved this by starting my prompt in ask mode in vscode and having it candidly plan changes so I can approve them. Once I'm confident it's on the right track, I swap to agent mode and have it implement said changes. Takes longer, but separating working tasks from conversations has been a better workflow overall

So, same concept for asking questions / discussing features. Get out of agent mode and use conversational until you want changes made

Re: Less human AI agents, please

#183

Post by anuramat » Sat, Apr 25, 2026, 4:24 PM UTC

disgruntledphd2 wrote: post #69

anuramat wrote: post #57

Earlier quoted context omitted.

they (are supposed to) produce average on average , and the output distribution is (supposed to be) conditioned on the context

Yeah but ultimately it's all just function approximation, which produces some kind of conditional average. There's no getting away from that, which is why it surprises me that we expect them to be good at science. They'll probably get really good at model approximation, as there's a clear reward signal, but in places where that feedback loop is not possible/very difficult then we shouldn't expect them to do well.

true, but it's the same with humans, we suck at problems with sparse/delayed feedback, which includes science (math would be the exception I guess)

sure, humans are obviously better at dealing with it, but the one thing nobody is claiming is "scientists replaced by 202X"

Re: Less human AI agents, please

#184

Post by jxf » Tue, Apr 28, 2026, 2:53 PM UTC

philipwhiuk wrote: post #80

jxf wrote: post #77

Earlier quoted context omitted.

It's not always amenable to grepping. But this is a great use case for AST searches, and is part of the reason that LSP tools should really be better integrated with agents.

Ah yes, don't fix the agents, fix the tools. What a ridiculously backwards approach. We were supposed to get agents who could use human tooling. Instead we are apparently told to write interfaces for this stumbling expensive mess to use. Maybe, just maybe, if the human can know to, and use, the AST tool fine, the problem is not the tool but the agent.

It's much harder to search using an AST tool for a human. It's certainly harder than grepping, for example. I use AST tools myself, but it takes a while to represent a complex structure in a big codebase when I need to look for that.