Caveat Prompter

Posted3 months agoActive3 months ago

azhenley

12 points

8 comments

surfingcomplexity.blogTechstory

calmpositive

Debate

20/100

LlmsPrompt EngineeringComplexity in Systems

Key topics

Llms

Prompt Engineering

Complexity in Systems

The 'Caveat Prompter' concept is discussed in the context of managing complexity, sparking a thoughtful discussion among HN users about its implications and applications.

Snapshot generated from the HN discussion

Discussion Activity

Light discussion

First comment

Peak period

132-144h

Avg / period

Key moments

01Story posted
Oct 13, 2025 at 8:00 AM EDT
3 months ago
Step 01
02First comment
Oct 19, 2025 at 3:50 AM EDT
6d after posting
Step 02
03Peak activity
5 comments in 132-144h
Hottest window of the conversation
Step 03
04Latest activity
Oct 19, 2025 at 1:23 PM EDT
3 months ago
Step 04

Generating AI Summary...

Analyzing up to 500 comments to identify key contributors and discussion patterns

Discussion (8 comments)

Showing 8 comments

fluxusars

3 months ago

1 reply

This is no different from reviewing code from actual humans: someone could have written great looking code with excellent test coverage and still have missed a crucial edge case or obvious requirement. In the case of humans, there's obvious limits and approaches to scaling up. With LLMs, who knows where they will go in the next couple of years.

greatgib

3 months ago

It is, because a human would have used "thinking" to create this piece of code. There could be error and mistakes but at least you know there there is a human logic behind and you just have to check for things that can easy mistakes for a human.

With AI, in the current state at least, there is no global logic involved with the whole thing that was created. It is a random set of probabilities that generated a somehow valid code. But there is not a global "view" about it that it makes sense.

So when reviewing, you will basically have to do in your head the same mental process as would have done an original human contributor to check that the whole thing makes sense in the first place.

Worst than that, when reviewing such change, you should imagine that the AI probably generated a few invalid versions on the code and randomly iterated until there is something that passed the "linter" definition of valid code.

satisfice

3 months ago

2 replies

In order to get the full benefit of AI we must apply it irresponsibly.

That’s what it boils down to.

stavros

3 months ago

1 reply

Which has also always been the same with people.

satisfice

3 months ago

1 reply

That’s what AI fanboys say every single time I make this point. But “it’s the same for humans” argument only works if you are referring to little children.

Indeed, my airline pilot brother once told me that a carefully supervised 7 year old could fly an airliner safely, as long as there was no in-flight emergency.

And indeed hiring children, who are not accountable for their behavior, does create a supervision problem that can easily exceed the value you may get, for many kinds of work.

I can’t trust AI the way I can trust qualified adults.

stavros

3 months ago

Well, you employ different adults than I do, then. Every person I know (including me) can be either thorough, or fast, as the post says, and there's no way to get both.

phinnaeus

3 months ago

At first I thought this was a typo, but actually I fully agree with this. If we use LLMs (in their current state) responsibly we won’t see much benefit, because the weight of that responsibility is roughly equivalent to the cost of doing the task without any assistance.

ofrzeta

3 months ago

Even that screenshot is bogus. When there's no understanding there can be no misunderstanding either. It's misleading to treat the LLM like there is understanding (and for the LLMs themselves to claim they do, although this anthropomorphization is part of their success). It's like asking the LLM "do you know about X?" It just makes no sense.

View full discussion on Hacker News

ID: 45567382Type: storyLast synced: 11/20/2025, 3:25:59 PM

Want the full context?

Jump to the original sources

Read the primary article or dive into the live Hacker News thread when you're ready.

Open link View on HN