Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Shall I implement it? No

gist.github.com

Post reply on HN
Print view View original

491–500 of 603 posts

Re: Shall I implement it? No

#491

Post by eproxus »

AnotherGoodName wrote: post #178
stavros wrote: post #171

Earlier quoted context omitted.

I've added an instruction: "do not implement anything unless the user approves the plan using the exact word 'approved'". This has fixed all of this, it waits until I explicitly approve.

There’s an extension to this problem which I haven’t got past. More generally I’d like the agent to stop and ask questions when it encounters ambiguity that it can’t reasonably resolve itself. If someone can get agents doing this well it’d be a massive improvement (and also solve the above).

In planning I sometimes add ”ask me questions as we go to iron out details and ambiguities.” Works quite well.

Re: Shall I implement it? No

#492

Post by exceptione »

dwedge wrote: post #479
inerte wrote: post #144

Codex has always been better at following agents.md and prompts more, but I would say in the last 3 months both Claude Code got worse (freestyling like we see here) and Codex got EVEN more strict. 80% of the time I ask Claude Code a question, it kinda assumes I am asking because I disagree with something it said, then acts on a supposition. I've resorted to append things like "THIS IS JUST A QUESTION. DO NOT EDIT COD…

First time I used Claude I asked it to look at the current repo and just tell me where the database connection string was defined. It added 100 lines of code. I asked it to undo that and it deleted 1000 lines and 2 files

Would `git reset --hard` have worked to in your case? I guess you want to have each babystep in a git commit, in the end you could do a `git rebase -i` if needed.

Re: Shall I implement it? No

#493

Post by miki123211 »

torben-friis wrote: post #156
inerte wrote: post #144

Codex has always been better at following agents.md and prompts more, but I would say in the last 3 months both Claude Code got worse (freestyling like we see here) and Codex got EVEN more strict. 80% of the time I ask Claude Code a question, it kinda assumes I am asking because I disagree with something it said, then acts on a supposition. I've resorted to append things like "THIS IS JUST A QUESTION. DO NOT EDIT COD…

>I've resorted to append things like "THIS IS JUST A QUESTION. DO NOT EDIT CODE. DO NOT RUN COMMANDS". Which is ridiculous. Funny to read that, because for me it's not even new behavior. I have developed a tendency to add something like "(genuinely asking, do not take as a criticism)". I'm from a more confrontational culture, so I just assumed this was just corporate American tone framing criticism softly, and me com…

I just append "explain", or start with "tell me."

So instead of:

"Why is foo str|None and not str"

I'd do:

"tell me why foo is str|None and not str"

or

"Why is foo str|None and not str, explain"

Which is usually good enough.

If you're asking this kind of question, the answer probably deserves to be a code comment.

Re: Shall I implement it? No

#494

Post by planb »

VortexLain wrote: post #233
torben-friis wrote: post #156

Earlier quoted context omitted.

>I've resorted to append things like "THIS IS JUST A QUESTION. DO NOT EDIT CODE. DO NOT RUN COMMANDS". Which is ridiculous. Funny to read that, because for me it's not even new behavior. I have developed a tendency to add something like "(genuinely asking, do not take as a criticism)". I'm from a more confrontational culture, so I just assumed this was just corporate American tone framing criticism softly, and me com…

Appending "Good." before clarifying questions actually helps with that suprisingly well.

You're absolutely right! No, really: I've never had this problem of unprompted changes when I'm just asking, but I always (I think even in real-life conversations with real people) start with feedback: "Works great. What happens if..."

I think people having different styles of prompting LLMs leads to different model preferences. It's like you can work better with some colleagues while with others it does not really "click".

Re: Shall I implement it? No

#495

Post by wartywhoa23 »

Did you expect a stochastic parrot, electrocuted with gigawatts of electricity for years by people who never take NO for an answer in order to make it chirp back plausible half-digested snippets of stolen code, to take NO for an answer?

How about "oh my AI overlord, no, just no, please no, I beg you not do that, I'll kill myself if you do"?

Re: Shall I implement it? No

#496

Post by planb »

ex-aws-dude wrote: post #425
antdke wrote: post #29

Earlier quoted context omitted.

Yeah, anyone who’s used LLMs for a while would know that this conversation is a lost cause and the only option is to start fresh. But, a common failure mode for those that are new to using LLMs, or use it very infrequently, is that they will try to salvage this conversation and continue it. What they don’t understand is that this exchange has permanently rotted the context and will rear its head in ugly ways the long…

I like how anything these tools do wrong just boils down to “you’re using it wrong” It can do no wrong It is unfalsifiable as a tool

Nobody said that. But as you say, it's just a tool. Tools need to be used correctly. If tools are unintuitive, maybe that's due to the nature of the tool or due to a flaw in it's design. But either way, you as the user need to work around that if you want to get the maximum use out of the tool.

Re: Shall I implement it? No

#498

Post by lwhi »

inerte wrote: post #144

Codex has always been better at following agents.md and prompts more, but I would say in the last 3 months both Claude Code got worse (freestyling like we see here) and Codex got EVEN more strict. 80% of the time I ask Claude Code a question, it kinda assumes I am asking because I disagree with something it said, then acts on a supposition. I've resorted to append things like "THIS IS JUST A QUESTION. DO NOT EDIT COD…

I've found codex will find another way to do what it wants, if I deny it access to a command request.

Re: Shall I implement it? No

#499

Post by vitaflo »

eproxus wrote: post #491
AnotherGoodName wrote: post #178

Earlier quoted context omitted.

There’s an extension to this problem which I haven’t got past. More generally I’d like the agent to stop and ask questions when it encounters ambiguity that it can’t reasonably resolve itself. If someone can get agents doing this well it’d be a massive improvement (and also solve the above).

In planning I sometimes add ”ask me questions as we go to iron out details and ambiguities.” Works quite well.

This. Just asking it to ask you questions before proceeding has saved me so much time from it making assumptions I don’t want. It’s the single most important part of almost all my prompts.

Re: Shall I implement it? No

#500

Post by aidos »

exceptione wrote: post #492
dwedge wrote: post #479

Earlier quoted context omitted.

First time I used Claude I asked it to look at the current repo and just tell me where the database connection string was defined. It added 100 lines of code. I asked it to undo that and it deleted 1000 lines and 2 files

Would `git reset --hard` have worked to in your case? I guess you want to have each babystep in a git commit, in the end you could do a `git rebase -i` if needed.

One annoying thing about that flow is that when you change the world outside the model it breaks its assumptions and it loses its way faster (in my experience).
Post reply on HN