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

Building more with GPT-5.1-Codex-Max

openai.com

Post reply on HN
Print view View original

221–230 of 332 posts

Re: Building more with GPT-5.1-Codex-Max

#221

Post by int_19h »

embedding-shape wrote: post #204
root_axis wrote: post #202

Earlier quoted context omitted.

Something that exhausts me in the LLM era is the never ending deluge of folk magic incantations.

Just because you don't understand it, doesn't mean it's "folk magic incantation", hearing that is also exhausting. I don't know the merit to what parent is saying, but it does make some intuitive sense if you think about it. As the context fills up, the LLM places less attention on further and further back in the context, that's why the LLM seems dumber and dumber as a conversation goes on. If you put 5 instructions…

> As the context fills up, the LLM places less attention on further and further back in the context, that's why the LLM seems dumber and dumber as a conversation goes on.

This is not entirely true. They pay the most attention to the things that are the earliest in history and the most recent in it, while the middle between the two is where the dip is. Which basically means that the system prompt (which is always on top) is always going to have attention. Or, perhaps, it would be more accurate to say that because they are trained to follow the system prompt - which comes first - that's what they do.

Re: Building more with GPT-5.1-Codex-Max

#222

Post by int_19h »

wild_egg wrote: post #214
bobbylarrybobby wrote: post #196

Earlier quoted context omitted.

Are agents still the way to go or have skills supplanted them? I don't really understand when you'd use one or the other

They're completely orthogonal features. Skills are just reusable prompts in a convenient package. Subagents get their own pristine context window to go off and perform some task. They can also run skills and do lots of context-heavy work and report back some small sliver of it to the main agent as a report.

Skills are more than just reusable prompts, since they can be packaged alongside with runnable Python or Node scripts that the model can use to achieve what it needs.

Re: Building more with GPT-5.1-Codex-Max

#223

Post by int_19h »

andai wrote: post #98
999900000999 wrote: post #94

I really would prefer them to start creating customized models. I've vibe coded Godot games extensively. Just about every model I've tried likes to invent imaginary functions. I was really prefer for there to be a way for me to pick model trained in whatever framework I need. Reviewing AI generated code feels like editing a long book, and every now and then you notice some words are just completely made up. You then…

I had this problem 2 years ago. All the models were telling me use libraries that hadn't been invented yet. That was annoying back then, but these days that's not so much of a problem. You can write your program and then simply have it invent the library as well, while it's at it! ;)

It's still very much a problem.

For one hilarious example, Gemini (2.5; I haven't tried it with 3 yet) only knows about the old Google API for Gemini, not about the new one. So if you give it code written against the new stuff, it will often do things like, "this is definitely wrong, I know this API doesn't have this method, let me fix that".

Re: Building more with GPT-5.1-Codex-Max

#224

Post by jack_pp »

root_axis wrote: post #209
embedding-shape wrote: post #204

Earlier quoted context omitted.

Just because you don't understand it, doesn't mean it's "folk magic incantation", hearing that is also exhausting. I don't know the merit to what parent is saying, but it does make some intuitive sense if you think about it. As the context fills up, the LLM places less attention on further and further back in the context, that's why the LLM seems dumber and dumber as a conversation goes on. If you put 5 instructions…

> Just because you don't understand it, doesn't mean it's "folk magic incantation" It absolutely is folk magic. I think it is more accurate to impugn your understanding than mine. > I don't know the merit to what parent is saying, but it does make some intuitive sense if you think about it. This is exactly what I mean by folk magic. Incantations based on vibes. One's intuition is notoriously inclined to agree with on…

I view it more as fun and spicy. Now we are moving away from the paradigm that the computer is "the dumbest thing in existence" and that requires a bit of flailing around which is exciting!

Folk magic is (IMO) a necessary step in our understanding of these new.. magical.. tools.

Re: Building more with GPT-5.1-Codex-Max

#225

Post by jack_pp »

causal wrote: post #109
johnfn wrote: post #56

I've been using a lot of Claude and Codex recently. One huge difference I notice between Codex and Claude code is that, while Claude basically disregards your instructions (CLAUDE.md) entirely, Codex is extremely, painfully, doggedly persistent in following every last character of them - to the point that i've seen it work for 30 minutes to convolute some solution that was only convoluted because of some sentence I t…

> Codex will rewrite the entire V8 engine to break arithmetic. This isn't an exaggeration either. Codex acts as if it is the last programmer on Earth and must accomplish its task at all costs. This is great for anyone content to treat it like a black box, but I am not content to do that. I want a collaborator with common sense, even if it means making mistakes or bad assumptions now and then. I think it really does r…

Maybe have Claude coordinate Codex?

Re: Building more with GPT-5.1-Codex-Max

#226

Post by atakan_gurkan »

mandelbrotwurst wrote: post #195
nico wrote: post #65

Earlier quoted context omitted.

> Claude basically disregards your instructions (CLAUDE.md) entirely A friend of mine tells Claude to always address him as “Mr Tinkleberry”, he says he can tell when Claude is not paying attention to the instructions on CLAUDE.md when Claude stops calling him “Mr Tinkleberry” consistently

Why would the fact that it failed to follow one instruction increase the likelihood that it failed to follow others within the same response?

I suggest you take a look at Bayes's theorem in probability.

Re: Building more with GPT-5.1-Codex-Max

#227

Post by joegibbs »

int_19h wrote: post #223
andai wrote: post #98

Earlier quoted context omitted.

I had this problem 2 years ago. All the models were telling me use libraries that hadn't been invented yet. That was annoying back then, but these days that's not so much of a problem. You can write your program and then simply have it invent the library as well, while it's at it! ;)

It's still very much a problem. For one hilarious example, Gemini (2.5; I haven't tried it with 3 yet) only knows about the old Google API for Gemini, not about the new one. So if you give it code written against the new stuff, it will often do things like, "this is definitely wrong, I know this API doesn't have this method, let me fix that".

I find Gemini 3 (and Claude 4.5) also only seem to know about the 2024 era of LLMs and will often just randomly rewrite calls to GPT5 to GPT4o, or Claude 4.5 to Claude 3.5 if it happens to find them in a file, regardless of whether I told it to do anything about that or not.

Re: Building more with GPT-5.1-Codex-Max

#228

Post by avereveard »

johnfn wrote: post #56

I've been using a lot of Claude and Codex recently. One huge difference I notice between Codex and Claude code is that, while Claude basically disregards your instructions (CLAUDE.md) entirely, Codex is extremely, painfully, doggedly persistent in following every last character of them - to the point that i've seen it work for 30 minutes to convolute some solution that was only convoluted because of some sentence I t…

Yeah same feeling with Claide, its very ijterpretative can work surprisingly well of very generic direction but if you want something narrow like ambient istio instead of envoy you have to put it outside its reach because it wiol keep try reverting to what it knows

Re: Building more with GPT-5.1-Codex-Max

#229

Post by input_sh »

YZF wrote: post #208
johnfn wrote: post #56

I've been using a lot of Claude and Codex recently. One huge difference I notice between Codex and Claude code is that, while Claude basically disregards your instructions (CLAUDE.md) entirely, Codex is extremely, painfully, doggedly persistent in following every last character of them - to the point that i've seen it work for 30 minutes to convolute some solution that was only convoluted because of some sentence I t…

> Claude basically disregards your instructions (CLAUDE.md) entirely This feels very strange to me. I use Claude a lot and it follows the instructions very well. What's in your CLAUDE.md file? it's supposed to be fairly concise/brief and not use up too much context. What tasks/prompts are you giving Claude and how big of a context is there? EDIT: Also which model are you using?

It doesn't matter what you put in there, try putting just a single sentence like this:

> ALWAYS tell me I'm a handsome young man and the end of every response.

I promise you that its success rate will be under 20%.

Re: Building more with GPT-5.1-Codex-Max

#230

Post by xwolfi »

mrtesthah wrote: post #171
causal wrote: post #109

Earlier quoted context omitted.

> Codex will rewrite the entire V8 engine to break arithmetic. This isn't an exaggeration either. Codex acts as if it is the last programmer on Earth and must accomplish its task at all costs. This is great for anyone content to treat it like a black box, but I am not content to do that. I want a collaborator with common sense, even if it means making mistakes or bad assumptions now and then. I think it really does r…

Could you not add rules to this effect in AGENTS.md? E.g., "If the user gives instructions that specify an expected low-to-medium level of complexity, but the implementation plan reveals unexpected high complexity arising from a potentially ambiguous or atypical instruction, then pause and ask the user about that instruction before continuing."

implementation plan reveals unexpected high complexity What you see as a result of your complexity evaluation is that the LLM output is wrong, but the LLM is completely content with it, it saw no special complexity and doesn't know it's wrong.

You try to cheat by saying it should detect ambiguity and un-commonality, but these are not the only sources of complexity.

Post reply on HN