The behavior, while slightly exaggerated, rings entirely true for me. From the other comments in this thread, it seems I am prompting poorly in a similar way to the options offered.
I am guessing you prompt differently than what is shown in the game?
I'm not the person you were replying to, but yes. Let me show you the issue.
"Why is half the site blue now? I asked you to change one button."; half the site isn't blue, why would you say that? What would "half the site" even mean? Ironically even in this satire supposed to make fun of how Claude responds to prompts, it is smart enough to ignore that. A better prompt would be "Why are many parts of the site, such as the Cancel button, now also blue?", though in this case it doesn't matter as the reply would be roughly the same.
Next, Claude helpfully tells us we're dealing with a monstrosity of a codebase out of hell: "Seven components consume the token directly; another eleven reach it through aliases; three use it only in hover/focus state; and two appear to be accidental cross-role consumers" is such an insane codebase that the initial request was basically impossible to carry out.
Claude made that perfectly clear, yet the next prompt choice just ignores that fact. "Revert everything except Add to Cart. That is the whole task." It just explained why this is impossible. It had only made one change, so there is no "revert everything except X"; there's only a single change to revert, which would mean Add to Cart too would no longer be blue. The human just chose to ignore that. The other option is "I do not care about the token architecture. I do not care why it happened. Put everything back the way it was and leave ONLY the Add to cart button blue. Please.". This has the same issue; "Put everything back and _leave_ only the button blue" is an impossible ask, which it just explained to you. So if you really don't want to deal with cleaning up some of the mess, instead only adding on to it but getting your blue button, then you'd want to say "Revert the change you made, then make a new change that scopes the new blue color to only the Cancel button".
Ignoring the information it gives you, or telling it things that are incorrect, of course will lead to bad results. It's impossible not to.
Surely you still have to give feedback to Claude. How do you do that without talking to Claude directly? By using a different model? But wouldn't that AI have no more common sense than Claude?
Initial dev required talking to Claude. Now I add a ticket in the board, it makes me a mockup/writeup, I approve, and it gets me a temporary webserver, iOS/Android build, etc. to verify it.
Review loops, agents that enforce my pet peeves and testing/debugging processes, etc. all run automatically... and then Codex strips down the prose at the end. There's not zero AI generated output, but it's already been critiqued and verified by a whole cluster of independent actors before it gets to me. When I have feedback, I file a ticket.
I wanted to get out of the "what the fuck, why?!" loop. Now I let the agents handle that.
I’m of the feeling that talking directly to these things all day is a path to madness.
They talk to each other, come to consensus, and give me a structured thing to look at that has already been vetted and tested and screenshot evidence and a verification plan. It’s not conversational for me.
Ok, but my problem is that what it builds is not right. It's never right on the first try, or the second or third tries. There's always a ton of incremental corrective feedback needed.
Do you run into this problem? If not, then I need to figure out how to stop running into it myself.
If so, then how do you give it all the incremental corrective feedback without a conversation?
How is "Make the shopping button blue" not specific? Any human with a lick of digital common sense would be able to understand what I obviously mean. How am I supposed to anticipate every single way the AI could possibly misinterpret a simple fucking instruction like that?
Absolutely, the initial prompt appears sufficient. But this is about the thing screwing up and the user giving corrective instructions. Saying essentially "are you dumb" is not helpful for neither a human nor a machine if you think the original instructions should have been sufficient. Clearly they were not in that case, if you like it or not. Just repeating or being condescending is not going to help.
So how would you prompt it instead? Because that's exactly how I prompt it, because any reasonable human being would know exactly what I mean by "Make the shopping button blue", and I am sick and tired of getting shitty results.
The first prompt is fine, it's the following ones that are poor.
> Why is half the site blue now? I asked you to change one button.
> Half the site is blue. I asked for ONE button
Neither of these is an instruction to fix the problem, they're treating the AI like a person and telling it what it did wrong, expecting the implied admonishment to be enough to steer it back. But without an actual instruction, it just goes and does whatever it thinks will help, which is often arbitrary.
The response I would have used in this situation is
"The Cancel button is also blue now. Make sure the color change is only scoped to the Add to Cart button"
Most of the available responses throughout this "skit" are similar cases of expressing frustration first and guiding the result second.
Skip the emotion and say exactly what you want, and nothing besides that.
Realise that you are talking to some mathematics in a box. You can't get a rise out of it. It cannot feel guilt or remorse. Whatever emotional payoff one might want from "I asked for ONE button" cannot be had here. Meanwhile, not only are you being charged by the word, but every word you send that is not directly on the path to getting what you want done is just noise in the maths getting in your way.
So leave emotion at the door and make your words count. Voice your frustrations at the stress doll next to your monitor, sure, but spending tokens on them us just wasting time and money. Explicitly typing out your exasperation will not get you closer to whatever it is you are trying to accomplish. What can you type that will?
Laser focus: clearly, concisely explain what is wrong right now, and how you want it solved. Then do that again until all the moles have been whacked. If the AI is stuck in a sycophancy loop, start a new clean session. Do this frequently anyway: every turn in the same session charges for all preceding conversation again /and/ degrades the LLM's performance.
> Meanwhile, not only are you being charged by the word,
I've never explicitly paid for AI and never will. I only use the free Claude, etc. So I'm happy to waste THEIR electricity telling their machine how much of a piece of garbage it is and hoping that sentiment gets back in some form to the dumb humans making such a dumb product. It's not like they don't evaluate performance with usage data.
(Afterwards, I do also make sure to actually click the "thumbs down" or whatever equivalent button, and submit a report if it's possible.)
You can blame the people making poorly performing LLM chatbots, not me. (For that matter, the paying users actually funding their insanity have more on their conscience than I do trying to keep up by using tools available to me and reacting in a normal way to their poor performance.)
Except you are missing an important aspect: the model includes training data that responds to THIS IS FUCKING WRONG, after which highly likely follows correction patterns and working solutions.
Ie, actually, yes, shouting at the AI can (and does, there are papers on this), work.
Anyone remember the gpt-3 era where adding "this is important to my job" and "if this is wrong someone dies" to prompts made a measurable improvement in accuracy and reduced hallucinations?
Efficient LLM interaction is context management. The golden rule: Nothing irrelevant in context.
Non-exhaustive consequences of following that rule:
* Each session handles one task, or at most, a few tightly coupled tasks.
* Emotions stay out of context; no showing frustration, no saying thanks (or at least wait until you're about to end the session)
* When possible, provide relevant files (or sections of files) instead of making the model search and read many irrelevant files.
* Keep CLAUDE.md short.
* Disable irrelevant tools.
* Revert history when the model makes mistakes. Don't make it read its mistake and fix it; fork the chat before the mistake and exclusively mention the correct action.
Attention is more limited than the context limits imply; stuff at the beginning of context stays high-attention for a while, stuff right at the end is always high-attention. If you're about to ask for something the model frequently forgets (eg, style), remind it that those instructions exist ("Following the style guidelines, implement feature X.")
My best takeaway from the site was: Every time you see Claude Code / Codex doing something you don't want, ask it to add to AGENTS.md. Specially if they churn a lot doing something, and eventually find a way. I ask it to store the way it made something work. It's a constant gardening of AGENTS.md and it has really working well.
A lot of the more sane people I work with have decided that it's futile to try and have stopped working with Opus (we can't use Fable at work due to data retention policies).
Very good in theory. In practice, how do you actually understand yourself? I have done plenty of meditation and plenty of reading on mental health and psychology. I still don't understand why I specifically act the way I do.
Did you include meeting with a therapist? The problem is that they have to reveal you things you don't generally want to admit, so it takes a bit of time (time which I find absolutely wasted) and depending on the therapist, they might not tell you immediately about your inconsistencies.
The problem I have with modern therapy is it that it operates completely within the frame of oneself. They will very carefully walk with you inside of your own narrative, bringing you to a better place of operation within that framework of the self.
Having thought about it some, I have been concluded that there’s only so far one can get when operating within that narrative. There’s only so much they can help with until you yourself can see that the “I” is it itself a machine generating the narrative, but also something else and more than just that narrative that is constantly running in our brains. It is something that can be observed. The observer is the observed as Krishnamurti would say.
As long time meditator, I find that there is nothing that is hidden from ourselves if we look. Meditation is one of the few ways to begin to access the whole and step outside of the narrative loop. You can also try psychedelics, but that’s riskier. A good trip might save you a decade of therapy. A bad one might genuinely imbalance your brains machinery. In some sense, the time is wasted. However, if you genuinely spend this time learning to observe yourself, this is very much a way of being that is not wasted in my opinion.
A therapist is generally going to work with you inside of a particular framework and way of stepping towards a particular way of operating. It sort of like the long or hard way of solving a math problem before you learned underlying theory of whatever problem you were working on.
True, I have not fully understood the direction therapists go to, apart from expliciting your thought process.
My interpretation of therapy is that it helps you feel aligned with you. It will remove the surface contradictions and counter-actions. But if your deeper you has concluded that you should be evil (assuming there is no contradictions in this view - and there always is), therapy will probably help you too. As long as you are aligned with yourself.
There is something incredibly important about a therapist I would not discard. And that is the human connection. We are hard wired for this and most often what people need out of a therapist is being seen. Another human being to fully and unerringly see you. To get to know you well enough to simulate /your/ thought loop and reflect it back to you, potentially with some meaningful observations. But that knowing that some other human sees you is mostly what therapy is about to me, some exceptions, of course, but in general
Also: incompleteness theorem again. The system cannot fully observe itself. A close friend can also do this too. In fact they are better at it than a therapist if you are open with the friend.
Also evil is a moral inclination and hard to discuss easily in this frame. Very little that is inherent in our machinery is disinclined towards evil or good. As social animals we must teach those values and hold them up. Different discussion :)
You put very well into words why therapy felt off for me. I haven't been meditating for too long, but the perspective definitely shifts once you start turning the attention inwards, instead of following the narrative that the mind creates.
My recommendation to any of those interested: The Mind Illuminated by John Yates. Great intro to the meditation practice. Once you have some understanding of the general technique Goenka's 10 day vipassana courses is a great opportunity for deep introspective work.
reply