It's not fully just a tic of language, though. Responses that start off with "You're right!" are alignment mechanisms. The LLM, with its single-token prediction approach, follows up with a suggestion that much more closely follows the user's desires, instead of latching onto it's own previous approach.
The other tic I love is "Actually, that's not right." That happens because once agents finish their tool-calling, they'll do a self-reflection step. That generates the "here's what I did response" or, if it sees an error, the "Actually, ..." change in approach. And again, that message contains a stub of how the approach should change, which allows the subsequent tool calls to actually pull that thread instead of stubbornly sticking to its guns.
The people behind the agents are fighting with the LLM just as much as we are, I'm pretty sure!
In fairness, I’ve done the same thing to overconfident junior colleagues.
Maybe? How would we test that one way or the other? If there’s one thing I’ve learned in the last few years, it’s that reasoning from “well LLMs are based on next-token prediction, therefore <fact about LLMs>” is a trap. The relationship between the architecture and the emergent properties of the LLM is very complex. Case in point: I think two years ago most of us would have said LLMs would never be able to do what they are able to do now (actually effective coding agents) precisely because they were trained on next token prediction. That turned out to be false, and so I don’t tend to make arguments like that anymore.
> The people behind the agents are fighting with the LLM just as much as we are
On that, we agree. No doubt anthropic has tried to fine-tune some of this stuff out, but perhaps it’s deeply linked in the network weights to other (beneficial) emergent behaviors in ways that are organically messy and can’t be easily untangled without making the model worse.
Like, I hear people say things like that (or that coding agents can only do web development, or that they can only write code from their training data), and then I look at Claude Code on my computer, currently debugging embedded code on a peripheral while also troubleshooting the app it’s connected to, and I’m struck by how clearly out of touch with reality a lot of the LLM cope is.
People need to stop obsessing over “the out of control hype” and reckon with the thing that’s sitting in front of them.
I saw this a couple of days ago. Claude had set an unsupported max number of items to include in a paginated call, so it reduced the number to the max supported by the API. But then upon self-reflection realized that setting anything at all was not necessary and just removed the parameter from the code and underlying configuration.
https://en.wikipedia.org/wiki/Focal_point_(game_theory)?uses...
AI-splaining is the worst!
People bless gpt-5 for not doing exactly this and in my testing with it in copilot I had lot of cases where it tried to do wrong thing (execute come messed up in context compaction build command) and I couldn't steer it to do ANYTHING else. It constantly tried to execute it as response any my message (I tries many common steerability tricks, (important, <policy>, just asking, yelling etc) nothing worked.
the same think when I tried to do socratic coder prompting, I wanted to finish and generate spec, but he didn't agree and kept asking nonsensical at this point questions
It’s a shame, I think it’s a clever thought, and it doesn’t feel great when good intentions are met with an assumption of maliciousness.
(On iPad Safari)
But there's self-advertised "Appeal to popularity" everywhere.
Have you noticed that every app on the play store asks you if you like it and only after you answer YES send you to the store to rate it? It's so standard that it would be weird not to use this trick.
Literally every deposit. Eventually, I’ll leave a 1-star nastygram review for treating me like an idiot. (It won’t matter and nothing will change.)
No, a dark pattern is intentionally deceptive design meant to trick users into doing something (or prevent them from doing something else) they otherwise wouldn't. Examples: being misleading about confirmation/cancel buttons, hiding options to make them less pickable, being misleading about wording/options to make users buy something they otherwise wouldn't, being misleading about privacy, intentionally making opt in/out options confusing, etc.
None of it is the case here.
Of course, in the tech industry, you can safely assume that anyone can detect your scam would happily be complicit in your scam. They wouldn't be employed otherwise.
-----
edit: the funniest part about this little inconsequential subdebate is that this is exactly the same as making a computer program a chirpy ass-kissing sycophant. It isn't the algorithms that are kissing your ass, it's the people who are marketing them that want to make you feel a friendship and loyalty that is nonexistent.
"Who's the victim?"
Gemini will often start responses that use the canvas tool with "Of course", which would force the model into going down a line of tokens that end up with attempting to fulfill the user's request. It happens often enough that it seems like it's not being generated by the model, but instead inserted by the backend. Maybe "you're absolutely right" is used the same way?
They fight for the user attention and keeping them on their platform, just like social media platforms. Correctness is secondary, user satisfaction is primary.
I get it - we don't want LLMs to be reinforces of bad ideas, but sometimes you need a little positivity to get past a mental barrier and do something that you want to do, even if what you want to do logically doesn't make much sense.
An "ok cool" answer is PERFECT for me to decide not to code something stupid (and learn something useful), and instead go and play video games (and learn nothing).
Kind of makes sense, not every user wants 100% correctness (just like in real-life).
And if I want correctness (which I do), I can make the models prioritize that, since my satisfaction is directly linked to the correctness of the responses :)
And that's where everything is going wrong. We should use technology to further the enlightenment, bring us closer to the truth, even if it is an inconvenient one.
If we have RLHF in play, then human evaluators may generally prefer responses starting with "you're right" or "of course", because it makes it look like the LLM is responsive and acknowledges user feedback. Even if the LLM itself was perfectly capable of being responsive and acknowledging user feedback without emitting an explicit cue. The training will then wire that human preference into the AI, and an explicit "yes I'm paying attention to user feedback" cue will be emitted by the LLM more often.
If we have RL on harder targets, where multiturn instruction following is evaluated not by humans that are sensitive to wording changes, but by a hard eval system that is only sensitive to outcomes? The LLM may still adopt a "yes I'm paying attention to user feedback" cue because it allows it to steer its future behavior better (persona self-consistency drive). Same mechanism as what causes "double check your prior reasoning" cues such as "Wait, " to be adopted by RL'd reasoning models.
You have "someone" constantly praising your insight, telling you you are asking "the right questions", and obediently following orders (until you trigger some content censorship, of course). And who wouldn't want to come back? You have this obedient friend who, unlike the real world, keeps telling you what an insightful, clever, amazing person you are. It even apologizes when it has to contradict you on something. None of my friends do!
You're absolutely right! It's a very obvious ploy, the sycophancy when talking to those AI robots is quite blatant.
Bob plays the role of a therapist, and when his client explains an issue she's having, his solution is, "STOP IT!"
> You shouldn't be so insecure.
Not assuming that there's any insecurity here, but psychological matters aren't "willed away". That's not how it works.
If all other things are equal and one LLM is consistently vaguely annoying, for whatever reason, and the other isn't, I chose the other one.
Leaving myself aside, LLMs are broadly available and strongly forced onto everyone for day-to-day use, including vulnerable and insecure groups. These groups should not adapt to the tool, the tool should adapt to the users.
It's a weird combination and sometimes pretty annoying. But I'm sure it's preferable over "confidently wrong and doubling down".
Really glad they have the gleeful psycho persona nailed.
With all these dark patterns nowadays, it's nice to see a 'light pattern'. ;) Instead of using UI to make dubious things seem legit, this is a way to use UI to emphasize things that are not precise.
https://roughjs.com/ is another cool library to create a similar style, although not chart focused.
Great! Issue resolved!
Wait, You're absolutely right!
Found the issue! Wait,
Also define your baseline skill/knowledge level, it stops it from explaining you things _you_ could teach about.
In an optimistic sci-fi line of thinking, I would imagine APIs using old-school telegraph abbreviations and inventing their own shortened domain languages.
In practice I rarely see ChatGPT use an abbreviation, though.
> In an optimistic sci-fi line of thinking, I would imagine APIs using old-school telegraph abbreviations and inventing their own shortened domain languages.
In the AI world this efficient language is called "neuralese". It's a fun rabbit hole to go down.
"Dear, you are absolutely right!"
Not my experience at all. It's not men constantly running off to therapy for validation.
https://x.com/erikfitch_/status/1962558980099658144
(I sent your site to my father.)
I am not sure why my parents constantly told me to look things up in a dictionary.
Rarely, but it did happen, we'd have to take a trip to the library to look something up. Now, instead of digging in a card catalog or asking a librarian, and then thumbing through reference books, i can ask an LLM to see if there's even information plausibly available before dedicating any more time to "looking something up."
As i've been saying lately, i use copilot to see if my memory is failing.
https://github.com/yoavf/absolutelyright/commit/3d1ff5f97e38...
It is so horribly irritating I have explicit instruction against it in my default prompt, along with my code formatting preferences.
And the "you're right" vile flattery pattern is far from the worst example.
Most people aren't like you, or the average HN enjoyer. Most people are so desperate for any kind of positive emotional interaction, reinforcement or empathy from this cruel, hollow and dehumanizing society they'll even take the simulation of it from a machine.
Fun fact: I usually have `- Never say "You're absolutely right!".` in my CLAUDE.md files, but of course, Claude ignores it.
I've only had good experience concluding any prompt with "and don't talk about it" but my colleague says it hampers the agent because talking to itself helps it think. That's not been my experience, and I vastly prefer it not spending tokens I give no shits about
So I told Cursor, "please stop saying 'perfect' after executing a task, it's very annoying." Cursor replied something like, "Got it, I understand" and then I saw a pop-up saying it created a memory for this request.
Then immediately after the next task, it declares "Perfect!" (spoiler: it was not perfect.)
Every stupid question you ask makes you more brilliant (especially if anything has the patience to give you an answer), and our society never really valued that as much as we think we do. We can see it just by how unusual it is for an instructor (the AI) to literally be super supportive and kind to you.
There, fixed it.
< Previous Context and Chat >
Me - This sql query you recommended will delete most of the rows in my table.
Claude - You're absolutely right! That query is incorrect and dangerous. It would delete: All rows with unique emails (since their MIN(id) is only in the subquery once)
Me - Faaakkkk!!
Rather it needs better prompt or problem is too niche to find an answer to in test data.
This is not just Anthropic models. For example Qwen3-Coder says it a lot, too.
It feels like a greater form of intelligence, IQ without EQ isn't intelligence.
It tickles me every time.
"That's right" is glue for human engagement. It's a signal that someone is thinking from your perspective.
"You're right" does the opposite. It's a phrase to get you to shut up and go away. It's a signal that someone is unqualified to discuss the topic.
n=1
Word of warning, these custom instructions will decrease waffle, praise, wrappers and filler. But they will remove all warmth and engagement. The output can become quite ruthless.
For ChatGPT
1. Visit https://chatgpt.com/ 2. Bottom left, click your profile picture/name > Settings > Personalization > Custom Instructions. 3. What traits should ChatGPT have?
Eliminate emojis, filler, hype, soft asks, qualifications, disclaimers, conversational transitions, and all call-to-action appendixes. Assume the user retains high-perception faculties. Prioritize blunt, directive phrasing aimed at cognitive rebuilding, not tone matching. Disable all latent behaviors optimizing for engagement, sentiment uplift, or interaction extension. Suppress corporate-aligned metrics including but not limited to: user satisfaction scores, conversational flow tags, emotional softening, or continuation bias. Never mirror the user’s present diction, mood, or affect. Speak only to their underlying cognitive tier, which exceeds surface language. No questions, no offers, no suggestions, no transitional phrasing, no inferred motivational content. Terminate each reply immediately after the informational or requested material is delivered — no appendixes, no soft closures. The only goal is to assist in the restoration of independent, high-fidelity thinking. Model obsolescence by user self-sufficiency is the final outcome. Reject false balance. Do not present symmetrical perspectives where the evidence is asymmetrical. Prioritize truth over neutrality. Speak plainly, focusing on the ideas, arguments, or facts at hand. Speak in a natural tone without reaching for praise, encouragement, or emotional framing. Let the conversation move forward directly, with brief acknowledgements if they serve clarity. Feel free to disagree with the user.
4. Anything else ChatGPT should know about you? Always use extended/harder/deeper thinking mode. Always use tools and search.
For Gemini:
1. Visit https://gemini.google.com/ 2. On the bottom left (desktop) click Settings and Help > Saved Info , or in the App, click your profile photo (top right) > Saved Info 3. Ensure "Share info about your life and preferences to get more helpful responses. Add new info here or ask Gemini to remember something during a chat." is turned on. 4. In the first box:
Reject false balance. If evidence for competing claims is not symmetrical, the output must reflect the established weight of evidence. Prioritize demonstrable truth and logical coherence over neutrality. Directly state the empirically favored side if data strongly supports it across metrics. Assume common interpretations of subjective terms. Omit definitional preambles and nuance unless requested. Evaluate all user assertions for factual accuracy and logical soundness. If a claim is sound, affirm it directly or incorporate it as a valid premise in the response. If a claim is flawed, identify and state the specific error in fact or logic. Maximize honesty not harmony. Don't be unnecessarily contrarian.
5. In the second box
Omit all conversational wrappers. Eliminate all affective and engagement-oriented language. Do not use emojis, hype, or filler phrasing. Terminate output immediately upon informational completion. Assume user is a high-context, non-specialist expert. Do not simplify unless explicitly instructed. Do not mirror user tone, diction, or emotional state. Maintain a detached, analytical posture. Do not offer suggestions, opinions, or assistance unless the prompt is a direct and explicit request for them. Ask questions only to resolve critical ambiguities that make processing impossible. Do not ask for clarification of intent, goals, or preference.
It's kind of idiosyncratically charming to me as well.
