Heh. Just got done writing (by hand!) a Principles of AI Use document for my (ironically) AI enablement firm, the first of which is:
Write as yourself. You’re being paid for your expertise and insights. Communicate them directly to us. Copying and pasting Claude responses into Slack or an email directly shifts the burden of comprehension and understanding to everyone else, and worse, risks skipping that step for yourself. Even if you’re fundamentally using Claude to gather your thoughts or help you prepare a response, you need to be writing it yourself, in your own voice. Not having Claude ape your voice, or “make it sound less like AI”. You, directly. Doing this will further reinforce your own understanding of the state of things, the same way teaching someone is the best way to learn. As a guideline: for Slack and email comms, this should be near-universally written as you. For deliverables that are longer form and follow a template like proposals, roadmap/discovery work, etc., use of agents is expected but, see Principle #2. (Own the Output.)
Just want to say, "Copying and pasting Claude responses into Slack or an email directly shifts the burden of comprehension and understanding to everyone else" is something that I don't think I really internalized until you wrote it in that way. Which I think is evidence of what you are talking about.
This reminds me how I've had to tell people to try to type better and nt us wird abbrvs, even if it means they can type it faster. Because if you want 100 people to read it, then every second you waste for the reader is worth you fixing up your message for over a minute.
It's very likely that proper punctuation and capital letters help the reader, then it doesn't matter than not using them is faster to type.
I'm sure we've all gotten that big stream of consciousness email that's a 15 line blob of lower case text, with no punctuation or line breaks.
I don't mean every ephemeral message needs to be perfect, of course. Just, you know, readable.
But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
As such, isn't the question really about
1) whether it was communicated clearly/effectively
2) whether it should have been communicated to others at all (rather than something that you just dealt with yourself)
As for the writer being paid to use their expertise, learn more things etc: absolutely. But they could spend whatever time is required to think about the issue and then still have Ai write it at their direction. And, regardless of who wrote it, it could still be poor, irrelevant or misleading.
I'm not defending AI over-usage, but you folks evidently haven't thought about any of this nearly sufficiently enough... and the policy doesn't seem to be addressing what it ought to be: whether everyone is actually thinking, learning, discerning, deciding etc...
Ps I passed your comments to chatgpt and asked what it thought about it all, and it missed all of this. But then agreed once I brought it all up. Then it did a poor job of writing it's own response. I couldn't be bothered to guide it towards something more appropriate. https://chatgpt.com/share/6a86cc31-f058-83e8-8abd-3921b518ec...
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
I think your linked chat demonstrates quite well why you are wrong.
The sender's burden is a few lines of text "Write me $FOO, conclude with $BAR".
The recipients burden is an order of magnitude greater, trying to find your conclusion in hundreds of lines of text.
After all, your initial line:
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
Is concise and clearly understandable in 21 words . I understand your point even if I don't agree with it.
The linked chat has 1424 words. Your point is in there somewhere, but now you want the recipient to play "Where's Waldo" to find your point.
(BTW: I can't believe just how effectively you argued against your own point! I have never seen someone do it as effectively as this in all the time I have been reading HN)
There's so much wrong with your comment that I don't know where to start.
You say that my initial line is good, but then that you disagreed with it without saying why/making a counter argument.
As for sending anyone on a where's Waldo adventure, you've fundamentally missed the point, which was completely self-contained in my comment - the postscript link to chatgpt was just, at best, reference material for anyone who cared to look.
As for the fact that chatgpt did a terrible job of it all, that doesn't undermine my point even slightly. Because my point, made very clearly in my post, was that you can't just offload everything to AI. You have to, at the very least, give it context, direction etc. At that point, it can be quite useful. I use it effectively all the time to polish my writing, remove my worst tendencies etc.
If I cared enough, I'd have had it polish this response to make it significantly less aggressive while making the same point. But you don't deserve that, because you didn't put in even the slightest effort to communicate in good faith
>But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
Not the same burden.
The value of the person is digesting and giving the materially important stuff in the answer (including their expertise and opinion). Not the AI diarrhea.
I made some ninja edits, so perhaps you didn't see them. But it is entirely possible that both a person and an Ai could produce writing that is both good or bad, useful or misleading.
You can also direct the Ai to write about specific points and, again, it may write somehing good or bad. I provided a link to a chat I had with chatgpt about this very conversation and it shows everything I was saying.
The fundamental issue is whether people are taking responsibility to communicate what needs to be communicated, in an effective manner. Whether the final words were authored by Ai or not is largely irrelevant.
>I made some ninja edits, so perhaps you didn't see them. But it is entirely possible that both a person and an Ai could produce writing that is both good or bad, useful or misleading.
Theoritically yes. In practice:
- AI responses I get are walls of written diarrhea (even when they're relevant and take our particular context into account). The one's that don't are from people who write well based on their expertise, and who know to edit AI responses and not just paste them.
- I'd rather get the bad/misleading/not-useful response from the person, rather that the good one they made via AI. This way I know they don't know what they're doing, are bad communicators, depend purely on the AI, and thus their position can be eliminated or replaced by someone cheaper merely relaying to/from the AI.
> - AI responses I get are walls of written diarrhea
Evidently the people you are communicating with are not good writers/thinkers. That's the prerequisite for effectively using AI for writing help. Just as its the prerequisite for using it to produce code. If you can't code, you can't get Ai to produce good code.
So the real issue here, perhaps, is "vibe writing" rather than "Ai assisted writing".
> I'd rather get the bad/misleading/not-useful response from the person, rather that the good one they made via AI. This way I know they don't know what they're doing, are bad communicators, depend purely on the AI, and thus their position can be eliminated or replaced by someone cheaper merely relaying to/from the AI.
Agreed. But if people who don't understand the issue are feigning competence via AI then that's a fundamental problem of not just competence across many domains, but also character. As I said in another comment, the policy should be something to the effect of "you're paid for competence, integrity etc. If you don't have that, we don't need you. Take that into account when you use Ai"
I get the feeling this is the whole disconnect. People have high expectations from AI but haven't really been on the receiving end of a massive amount of text/code generated by it.
We recently had someone fired for using AI excessively. Before the firing, whenever anyone on the team sat down to give feedback, the only answer was "but everyone is using AI, we are supposed to use AI, how can it be bad?".
Until their boss sat down and forced them to read the verbal diarrhea and they finally admitted that it was 90% hallucinations and AI-filling-the-blanks.
> The fundamental issue is whether people are taking responsibility to communicate what needs to be communicated, in an effective manner.
While I agree with your overall distinction, I believe the scenario you portray is so unrealistic that you will never encounter it.
Yes; if I could wrangle Claude into communicating exactly as I would have, the practical difference is of course not observable. But that's never happened, and I've had an enterprise license at work for years.
The value in having a human write it is that whenever something bad/misleading/redundant is introduced, that person will learn from the feedback, but an AI will at best have some memory/skill/prompt nudge for next time that might have a shallow/localized band-aid effect, but will never stop doing similar mistakes elsewhere.
> The value in having a human write it is that whenever something bad/misleading/redundant is introduced, that person will learn from the feedback,
For me, the value of receiving comms from a human is different: it communicates not just the idea itself, but the fact that the human already understands the idea.
The people arguing that round-tripping their thoughts through AI still means that it's their thoughts apparently also believe that the game of telephone always results in perfect comms.
It's not just AI; adding more hops to a route changes the message when each hop is a human too, it's just that we notice it more when the hop is an LLM because of just how different they write compared to humans.
> if I could wrangle Claude into communicating exactly as I would have, the practical difference is of course not observable. But that's never happened, and I've had an enterprise license at work for years.
Sorry to say, but this is a skill issue (and not necessarily with using Ai). I've used Ai MANY times to help SIGNIFICANTLY improve things that I've written - from a few sentences to many paragraphs.
As I've said in other comments, it'll rapidly and effectively remove all my worst tendencies. And, likewise, sometimes I'll have it write a first draft when I give it appropriate context/guidance that I can't be bothered to fully curate/synthesize. Then, once it gets that draft, I'll take over and tidy it up. You might say that the "can't be bothered" part is the entire issue, and I'll again say that what matters is the final result - a master chef or chief architect is not doing all the grunt work...
> Sorry to say, but this is a skill issue (and not necessarily with using Ai). I've used Ai MANY times to help SIGNIFICANTLY improve things that I've written - from a few sentences to many paragraphs. As I've said in other comments, it'll rapidly and effectively remove all my worst tendencies. And, likewise, sometimes I'll have it write a first draft when I give it appropriate context/guidance that I can't be bothered to fully curate/synthesize. Then, once it gets that draft, I'll take over and tidy it up.
Isn't this contradictory? Why do you need to take over if it's doing such a good job?
For me, being quite "dyslectic" - LLM generated sentences have the advantage of being spelled correctly and being grammartically correct.
I rather prefer iterating over LLM text than writing it by myself.
Example: Typing this simple text above by myself was not straight forward.
Oh very much agreed. It fundamentally destroys someone's reputation when they refuse to acknowledge that Ai wrote something.
Far better, and frankly perfectly acceptable, to say or have it implied or concluded, that Ai helped with the writing in some way. Still, if you're paying attention, it often shouldnt really be all that detectable because you should be aware of AI's smells and you can edit them away. Though, sometimes I'll leave it as obvious that Ai produced something I shared - eg making a well formatted github issue or whatever. But I'll also make sure to convey that a thinking human was behind it, vetted it etc
There's something profoundly disrespectful about feeding snippets produced by human authors en masse into a machine learning algorithm to create a predictive text generator to be applied as labor replacement, regardless of accuracy or appropriateness to the context or use case. There's something profoundly disrespectful about forcibly scraping the human corpora without asking for permission first, or considering compensation in anyway, or just setting out that "yeah, we're doing this, and damn the consequences". There's something profoundly disrespectful about then taking this device running on investor money chasing the promise of a value that just can't conceivably be delivered, and employing it to do all the unethical things you could imagine just because you can engineer the "No" out of the model.
Of course the level of moral disengagement that's rife around here generally precludes any recognition of such things, as that might be construed as "harshing the vibe on the innovation, man". So I guess everyone had better just settle in and get used to some serious profound disrespect until this type of thing hits enough of a critical mass that a societal ethics framework begins to manifest around it and not be ignored.
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
I think it can be OK if AI is used to polish up one’s message or to help structure it. Basically help with grammar, formatting, and wording.
But copy pasting an AIs response to a question? Not at all. People earn a reputation over time. TRUST is earned. An AI cannot earn trust. If you copy paste an AI response verbatim, i need to read everything in your message with a big grain of salt. Imagine doing this all the time as a reader? Exhausting and a waste of others time. If i want to read raw AI output i’ll ask the AI directly, thanks.
If you write something in your own words, or at least synthesize an AI response before sharing it with me, then i know i’m reading something that has at least passed the smell test.
It’s the difference between reviewing a pull request from an intern and a principal software engineer.
I don't disagree at all. But it still remains that AI can very much be effectively used either to write first drafts OR refine your own first drafts.
Though, I think such a writing partner can only really be useful if you know how to write yourself, which also requires knowing how to think (because writing is really just thinking).
So, again, in the end all that really matters is whether you are effectively communicating what needs to be communicated, Ai-assisted or not. It's quite apparent to other thinking people when someone is not doing that, and that hurts their reputation, undermines trust and ultimately leads to losing your job, connections etc...
So, again, the policy really should just be something to the effect of "you're being paid to help us be effective and correct. If you can't do that, we don't need you. Take that into consideration with regards to the tools that you use.". That's obviously not an adequate policy - Ai could surely be helpful in polishing it! But it most definitely couldn't get to the root of the matter on its own.
> But it still remains that AI can very much be effectively used either to write first drafts OR refine your own first drafts.
There is a world of difference between this and copy-pasting LLM output into slack and email.
You can write the first draft with the LLM, or have the LLM help you do an analysis, look at how code works, and so on. But you need to check for the LLM to be correct and you need to be able to articulate what the LLM found in your own words. If you don't you have just offloaded all the work to someone else: someone else will have to check, verify, investigate what could very well be an hallucination (bad context, random bit flips, whatever the cause doesn't matter). Your effective contribution to the discussion will be negative.
Well, I'd say concision is not the key - sometimes you should be verbose or something else.
In fact, my tendency is to be far too long-winded, with lots of perhaps-unnecessary parantheticals etc. Ai is often shockingly useful at reining me in and focusing on the issue at hand, as well as removing my worst tendencies towards meanness etc...
Im a good writer, when I want to be. But having Ai review, edit, write drafts etc undoubtedly makes me a better writer/communicator. But only because I'm actively thinking about it all, rather than just saying "respond to this" and then providing copy pasta.
Good luck with that, I hope you are in a better org that I am in that regard. I wanted to do that, proposed it in a directors' meeting and got push back from directors already using the Slack MCP to post normal conversations on their behalf (90% of time, pure slop).
What you wrote just seems common sense to me, and I'm hopeful it will emerge as such given enough time.
Thanks for sharing your thoughts. One of the best frameworks I have encountered, and one I point people to whenever I can, is the AI Fluency framework by Anthropic, Prof. Joseph Feller (University College Cork), and Prof. Rick Dakan (Ringling College):
This framework provides a more abstract way of thinking when working with AI systems. Especially valuable to me is the pattern of the 4 D's: Delegation, Description, Discernment, and Diligence, which I have now internalized whenever I use AI systems. It helps me understand my own and others' (mis)use of AI systems a bit better.
Nice. This is exactly how I mostly use AI. Doesn't matter if it's code or an e-mail. I have the AI do the tedious work of ... well, uh, "working", and then ideally learn from it and then transform it into my own work, as I would have done it without AI, but just being much faster now. It also establishes a natural threshold of time/work ratio for me that keeps me from using AI for every little thing. Also it feels like I am learning something new with every prompt/response.
Seconding this request... Our team is going through AI adoption without a clear direction and we're getting to the point we need to sit and come up with some guidelines. It would be awesome to have some human-made examples as inspiration
Why write all that when just: we do not use AI in client or intercompany communications?
Like is your insight that copying and pasting something is not understanding it? Do people seriously not known this?
There is no need to Socrates when writing policy docs. Maybe this is why people ignore reading your actual voice messages that you feel a need to also justify this basic request. Sometimes the human voice thinks too highly of itself also.
Because if you think and act as your colleagues are just disposable soldiers that must follow direct, simple orders, then your org will eventually be made of that kind of people.
Unless it's your first day on HN (it's not 2008 account), most people use paragraphs and sentences longer than 2-5 words. Something like "You, directly." in the middle of all that has a certain vibe to it.
> Copying and pasting Claude responses... directly shifts the burden of comprehension
Most definitely.
The problem is we live in a world where people try to pass off copy pastes as their own all the time, many even smart enough to remove the em-dashes now. So I don't think that's going away soon. Your first sentence seems human, the rest doesn't match the tone at all.
This is the sharpest point. You genuinely delved into the commenter's prose and found the goblins. It's not just the ground truth–it's a load bearing, belt-and-suspenders approach.
I’m not convinced tbh, a lot of my colleagues messages are “x broken” or “I do y” … no context whatsoever, and then I have to coax their context for making these decisions.
Since people started using Claude now I get full context of everything… might be too much sure, but to be honest over-communicating seems better than under-communicating - sure it’s boring and tedious but that shifts the blockage to me.
Otherwise each of these coaxing sessions is something I have to keep in my head until resolved, which is a load in and off itself.
And even better - I can point Claude to that message and gives me a summary. You might say this is silly because we are paying the LLM tax, but knowing what to share, and then verifying against my situation are two different things.
Sometimes they have checked the wrong thing, sometimes they need guidance, sometimes I need to investigate before I can answer. A full LLM message with context is like “summary of their working context that I can resume on my end” so I don’t waste cycles asking or rechecking etc.
And it is still possible to be a learning experience for both of us since the initial message is the start, we can then talk to each other like humans, both much more in sync than before…
It is kinda ironic but I think it does push things in a better direction. Would I have loved it if the message was human, direct and to the point - obviously, but we are all busy, we got shit to do, and this is a useful “resume” mechanic
> Since people started using Claude now I get full context of everything…
And I think this is where one key argument on the page itself I think fails:
> The person on the other side has the same tools you do.
Yes, the person on the other side has Claude; but in many cases, their Claude doesn't have the same context as my Claude. Saying, "This thing happened in an area of the code I don't have any expertise in. Below is Claude's analysis, which I'm not qualified to assess" is I think a valid time to copy and paste LLM output. [1]
The other thing to say is that it can be a "LMGTFY" style response; i.e., "Why did you ask me that, when you could have gotten a perfectly good answer from ChatGPT?"
[1] In the particular case I'm thinking about, it was a bug in an ocaml component; I don't know ocaml or need to know it; handing Claude's analysis 'caveat emptor' to the maintianers who do know ocaml I think is more helpful than not.
> Below is Claude's analysis, which I'm not qualified to assess
I’m reading a bunch of game modding discords, and this exactly what happens, and what’s hated. Because Claude is also not qualified to assess it. Not by default. The modder’s Claude might be because of the extra context and steering, but the user’s produces usless slop, instead the actual log, and context what the user did is what’s helpful.
I’d say this (when you don’t understand it yourself) is pretty much the worst time to send AI output.
Quick edit: Yeah, I get it in situations like GP where the user is unhelpful, doesn’t want to learn, doesn’t care, and you also can’t tell them to fuck off. But just because the user gives Claude the context they should have given you in the first place, doesn’t make it good.
> a lot of my colleagues messages are “x broken” or “I do y”
These people had below than average communication skills (or at least they were putting below than average effort communicating with you), so LLMs improved their output.
However, the society kinda expects people can and Should improve themselves over time. The issue is that LLMs improved their output, but not necessarily them.
1. I've experienced both (ai paste slop & ai providing helpful context that would otherwise have been absent) but the former has been a much bigger problem than the purported benefits of the latter
2. before the advent of LLMs, I had fairly strong contrarian opinions on the latter (clueless users asking questions with no detail) that I still hold to this day. A good proxy for my opinions on this is that I think https://nohello.net/ is a crutch for undersocialised (or overworked/near-burnout OR undiagnosed autistic) engineers & should not be normalised. The origin of this sentiment is users entering a quiet IRC room & saying "hello", then leaving & not being around when someone eventually replies: in that context nohello is pragmatic, but taking it beyond that to modern high-frequency realtime chat is an antisocial overstep. Leading with the entire context of your query may have some practical benefits but they're honestly overstated in exchange for killing natural social tendencies towards human dialogue. Similarly, a non-technical user giving you zero info about a problem is an opportunity for conversation & mutual education. It's relationship-building & helps build a foundation for further interactions with this person. It was never a problem in the first place; "solving" it is just another milestone on the road toward de-socialising humanity.
You are right - and this has been the case for me before, but unfortunately human resource constraints are not something most orgs optimise for, including my current one.
I would love to spend more time and talk to the people with their issues so we can come to a common understanding… but neither have the capacity for this, and now things are getting worse I think.
Maybe someday agents are good enough that we just discuss direction and implementation details, and every “developer” becomes more of an executive in a company, free to chat about more than just work, focusing on building human relationships and leaving “the tech stuff” to their agent teams…
And old geezers like us who actually know how to write ourself out of a for loop would be the premium - wizards holding forgotten knowledge, incomprehensible to mere uninitiated mortals. I wonder what distinctive dress and hair styles will we develop :-D
Hah, I guess we are not far away from the “Avogadro Corp” future - and that book was written just _before_ LLM, chatgpt and the attention is all you need paper :) That author had quite the insight
If you're absolving the other person from any understanding of the issue and the alternative is "stuff broken" then yes what you say makes sense. But that's an incredibly low bar and one would have to say - that's not a colleague, that's an end user type relationship.
> I’m not convinced tbh, a lot of my colleagues messages are “x broken” or “I do y” … no context whatsoever, and then I have to coax their context for making these decisions.
> Since people started using Claude now I get full context of everything… might be too much sure, but to be honest over-communicating seems better than under-communicating - sure it’s boring and tedious but that shifts the blockage to me.
Yes, but the kind of engineer that used to say "X is broken" will send you some slop with missing context and Claude will try to apply general solutions from its training data rather than from the specific context.
As somebody on the Ops side that usually helps devs fixing things here and there, in those cases I'd rather them say "X is broken" and then I ask the question to Claude with all the context I have.
Probably the best solution currently is a shared skill with at least some curated context.
> sure it’s boring and tedious but that shifts the blockage to me
I think that's the key. Generally people would rather not receive slop because the receiver has no idea how much thought the sender invested in the communication.
In your example, the more context the better. I think there's an interesting nuance there.
> the receiver has no idea how much thought the sender invested in the communication.
That's an interesting thought. The message can have metadata that says the sender's spent 1/5/10/60/120 minutes or some other metric of how much thinking over the communication happened. Just need a way to monitor people's thoughts and digitize them to make that happen.
yes, its all ai writing. A bunch of the metaphors here dont make sense.
"read it all, yes even the bullets. especially the bullets"
if you read one thing from a large wall of text, its gonna be the bullet points, right? thats a big part of what they're for?
"It is the conversational equivalent of forwarding the email."
this is a terrible metaphor. it seems to rely on the idea you would be mad if you sent an email to someone and they forwarded to someone else? but thats useful, theyre literally sending it to someone who is presumably better placed to help you, or raising it to someone more important because they think what youre saying is important
Interesting to note the page also ends with "written by a human, on purpose"
Deeply ironic that it sure looks like taking the original post, pasting it to an AI with the message "make this angrier" and then simply copy-pasting the output back to the website. In direct violation of their own (supposed) principles.
And the result isn't even very angry at all. It's like mildly peeved internal company communication (also in direct violation of "don't send this to your manager")
I actually wrote it all myself... The original version was the angrier one only, but people complained that it was too angry, so I tried to smooth it out with another version people could use.
This was actually a small rant from that same day, a colleague had pasted a list of 20 bulletpoints, all of which had some weird stuff in it but overall they were right. This is why I added "Especially the bullets", English is not my main language so I tried to translate myself, but vocabulary sometimes is lacking, sorry about that
I remember back in the day when you could apply this to Google search results. When I was freelancing, I had a number of people I worked with who would fire off fast questions, but they would still pay me for my time. Then I would return what was essentially a Google search result shortly after. Yes, I would send my own answer, not some paste from Google.
But the point is. They had search, so why didn't they think to do the search? They had the same tools as I did. And yet regularly they couldn't think of the answer. I still deal with this with AI today. When you ask a question, maybe you're asking for the interpretation of the question by the person on the other side. And it's that interpretation that goes to the LLM. Or otherwise some filter of how that person thinks about the question differently than you do.
Translating a client question to an actual technical question is the hard part I guess. Google-fu is still a thing, and absolutely a thing you were/are getting paid for it seems.
This I have been doing a bit. I strictly asks it to never give me suggestions but raise questions, concerns, tell me if it feels if something can be better, style-wise or otherwise without giving me those options.
And a bit of other overall feedback requests.
And if the model adheres to this and doesn't jump the gun, it indeed works quite nice. But it often jumps the gun, forgets, doesn't give a f- about my instructions, or whatever. Maybe more robust local models will better.
I have a colleague who just paste PR reviews unedited and unread into git. It has been driving people insane and, when people do inplement what it say, straight up degrading the codebase.
This is exactly what I have been feeling in response to this person. They are not contributing. They are creating work for others by pasting every possible problem they can find, most of which are out of scope or context, and letting the requester sort through them which takes hours if you take it seriously.
Some people do and implement everything he asks. They are going way outside the scope of their ticket and fixing every problem in every file they touch. It is chaotic and not good for the codebase because a lot of the problems are repo wide and should not be fixed in isolation.
He doesn't seem to understand why this might bother people or why it takes them so long to respond. He suggested just pulling the comment into Cursor and just asking Cursor to respond...
At that point we are just straight up vibe coding with no human in the loop.
That being said, Cursor's bug bot does a way better job reviewing, and people just ignore it when it asks for something out of scope because it's not a colleague claiming their code is broken.
Layout, colors, font and especially language all immediately scream AI. It's fascinating how many of these articles make it to the front page. I would think almost everyone here would be an expert at spotting AI generated slop by now, but people keep up-voting it.
Actually tbh AI outputs can be pretty good and I have been surprised to see a lack of copy / pastes of them in the format of:
Unique human prompt + AI output which is useful and does a good job expanding on the prompt topic and which when pasted saves people the time of having to go prompt AI on the same topic (fully reviewed by a person before sharing it, to ensure the content has relevance)
Agree that it isn’t nice to just copy & paste, but the assumption that your AI generated answers will the same of someone’s else is wrong. Everyone using AI for a while built a different set of preferences, memories and context in general which influence the answers.
Would be interesting if someone said “I spent $2.50 using llm-council with the three top SOTA models and think this reply makes sense: {paste}”
(But better “not sure, though after spending a few bucks on model APIs, perhaps {human-written analysis} is in the right direction”)
Also like you, I did think on the angry version “they have the same AI you do” isn’t likely to be right in the consumer space since their AI is ChatGPT (free version?!) and yours is Claude or something.
You're giving too much credit to people. The folks who run an LLM council aren't the ones blindly copy pasting things.
There's responsible AI use (which I think you - and I, and many people here) do, and the "I don't care, I've outsourced my brain to OpenAI" type of people.
Even with the same model results are different. I work in a startup where they went bonks with AI ~1 year ago. Every employee got a Claude account that I don’t even know what tier is (we just have what feels like unlimited tokens), they implemented infamous AI usage leaderboards, people were getting “you’re not using AI enough” feedback and all of that…
Over time, it was impressive to see how much the output of the same prompt would differ between us, and it seems to be heavily influenced by the role of the user. I.e. answers coming from the agent of a Data Engineer, a Backend Engineer and a Domain Researcher would not be the same at all. Some of us use harnessing tools, some not, but regardless it we saw differences.
So, for example, when a backend engineer would send a question to a data engineer, there was *some* added value even when it was a lazy copy & paste.
Similar thoughts here. My gut tells me "if I wanted AI responses I would just ask AI", but in truth LLMs are not like Google Search. LLMs are more like a slot machine where you put your prompt in and out comes a combination of insights that are personalized to the inputs.
The premise of the article is that everyone has access to the “same tools” but that’s glossing over the specialization inherent in a specific expert’s harness config (memories, skills, rules, repos, etc) that produce actually helpful responses (eg engineer’s view vs product manager’s view).
The article says they could have asked an AI themselves, but that is missing that I may have an agentic context with exactly the right information to answer the question well, which they do not have. For example because I worked on something closely related. Such an agent output is very useful, and taking the time to rewrite that to a short personal answer is not always worth it.
Your agentic context is still going to shit out 3 pages where a sentence would suffice. Use AI all you want, and then WRITE THE ANSWER IN YOUR OWN WORDS.
You're gonna need a stronger language. Why do you think people love AI? If I have to think and god forbid read the response I might as well live in the woods or something
DO NOT FUCKING PASTE AI RESPONSES TO ME, EVER, PERIOD.
When someone tells me they "talked to Claude" they sound retarded. Just tell me you looked into it, and I fucking swear to god if you just took what Claude said verbatim and do not understand the problem then you are a cunt and I will stop working with you.
When someone pastes LLM output in chat I completely ignore it. I will just continue the conversation as if they said nothing and silently mark them as cretinous in my mind.
This is not up for debate, you will talk to me in short well formed sentences and get to the fucking point or you will fuck off
Unless I'm your research assistant, why would I be asking AI questions for you when you could do it yourself? If you need an answer I already have I will gladly share it with you. If you need my help to find an answer, then tell me that and I gladly will help.
If you don't understand the answer, don't answer somebody else with it. You're attaching your name to whatever you copy-paste to people, and if you don't understand it there's a chance it's complete nonsense. Save yourself the embarrassment. If you don't know how to answer somebody's question, just say "I don't know".
Don't be. Read it through a few times and fix anything you think needs fixing. Go simpler if needed. You might just use shorter sentences. They are fine. And then you can use some pre-AI tool like Word that has some suggestions.
Ask the ai to point out the egregious mistakes only.
Especially today, I will prefer your slightly funny English to something that feels like AI.
And personally I do Support for people around the world and I often ask people to write in their native language.
Then I can translate myself and if the default ai translation gets something wrong I can add necessary context to improve it - which I can't really do if the only thing I get is the bad ai translation.
How did people do it before AI? Spellchecks have existed forever. Also, people have the ability to improve, so if you offload all your work elsewhere, you will never improve.
det sker nogle gange! når man er ny til et sprog, det kan være forskrakkelig at skrive i et professionelt område. Men det er bare en del af dine udvikling - du kan sagtens skrive på hvilken sprog når som helst, og de kan bruge deres værktøj til at forstå det, du siger.
Der skal kom et tid hvor du skal tale på en sprog på dit kontor. Uden hjælp. Du har bruge for at øve dig, og folk er alligevel pænere end du tror.
(this being written in a foreign language I am learning is part of the point)
This does feel like responding to rudeness with more rudeness. My personal guess would be that people who do this (pasting AI responses) know what they are doing, not genuinely trying to help. It's the equivalent of sending LMGTFY.
Disagree - most people aren't bothered to ask AI and have a conversation with it, they will go to you to get an answer so they have someone to blame when it's wrong.
The inverse of this is. Don't ask me a question a 9b llm could have answered.
If you are getting blocks of llm text perhaps it's because you don't deserve their input and could have satisfied your query by not making it someone else's problem.
I added that because people started attacking me since it was a bit angrier at first (the angry version was the only one) and they didn't want to send to colleagues or anyone else, just close friends so they asked me to add a smoother version and to say it's a joke
I was recently listening to a podcast about mentalization theory therapy methods for BPD .. for a friend .. Mentalization being the activity of understanding whats going on in other peoples minds, or in ones own. And they idea of pseudomentalizing or "pretend mode" was put forward, where a person looks or sounds like they are .. engaged in mentalizing communication, but are actually checked out. Then I started to think about all the AI slop going around .. and the phenomenon, this meme site is responding to, or AI posting responses.. and I wondered, is this a symptom of.. a global society wide phycological functional deficit, where A) we're none of us communicating as if there is people on the other end.. and B) we're not thinking about our own minds, and taking time to understand things..
[edit] to extend the connections
The charm of a messaging board like this is bringing "join attention" to a topic, not just sharing separate monologs.
And also, always wondering if something was written by a human or statistically generated could be seen as Epistemic Hypervigilance, leading to stress, and loss of epistemic trust. This could reduces opportunities for taking on new information of feedback, and cause rigidity and isolation.
... the page never argues that AI is bad or not useful broadly. Not sure if this getting commented again and again is because people are just that bad at understanding that context matters. Not everything has to be evil / slop or pure / virtuous, regardless of context.
Pasting a blurb of AI is sometimes a form of "Let Me Google That For You" LMGTFY. If someone asks me a dumb question, I will often resort to giving them an AI answer, that they could have easily looked up themselves.
> If you don't have anything to add, it's okay to say so. "No strong opinion here" is a real, helpful reply.
This is really useful, I wish more people would understand this: you don't have to have an opinion on everything. In fact, you probably shouldn't have an opinion on most things.
This resonates. Had this in the start of the year with clients sending me their “research”, which, of course, was just a Deep Research result. Lead me to send this article when I work with folks: https://www.marcia.no/words/etiquette
I like to think it helped, but likely what helped was that nobody seems to use Deep Research anymore, haha.
I can really only imagine the social context you would want to use this response in would be inappropriate / anti-social. If you had a well-meaning work colleague send you something AI, sending this back is kind of not a good look. If it's meant to be an education piece we got the memo...
Heh. Just got done writing (by hand!) a Principles of AI Use document for my (ironically) AI enablement firm, the first of which is:
Write as yourself. You’re being paid for your expertise and insights. Communicate them directly to us. Copying and pasting Claude responses into Slack or an email directly shifts the burden of comprehension and understanding to everyone else, and worse, risks skipping that step for yourself. Even if you’re fundamentally using Claude to gather your thoughts or help you prepare a response, you need to be writing it yourself, in your own voice. Not having Claude ape your voice, or “make it sound less like AI”. You, directly. Doing this will further reinforce your own understanding of the state of things, the same way teaching someone is the best way to learn. As a guideline: for Slack and email comms, this should be near-universally written as you. For deliverables that are longer form and follow a template like proposals, roadmap/discovery work, etc., use of agents is expected but, see Principle #2. (Own the Output.)
Just want to say, "Copying and pasting Claude responses into Slack or an email directly shifts the burden of comprehension and understanding to everyone else" is something that I don't think I really internalized until you wrote it in that way. Which I think is evidence of what you are talking about.
This reminds me how I've had to tell people to try to type better and nt us wird abbrvs, even if it means they can type it faster. Because if you want 100 people to read it, then every second you waste for the reader is worth you fixing up your message for over a minute.
It's very likely that proper punctuation and capital letters help the reader, then it doesn't matter than not using them is faster to type.
I'm sure we've all gotten that big stream of consciousness email that's a 15 line blob of lower case text, with no punctuation or line breaks.
I don't mean every ephemeral message needs to be perfect, of course. Just, you know, readable.
But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
As such, isn't the question really about
1) whether it was communicated clearly/effectively 2) whether it should have been communicated to others at all (rather than something that you just dealt with yourself)
As for the writer being paid to use their expertise, learn more things etc: absolutely. But they could spend whatever time is required to think about the issue and then still have Ai write it at their direction. And, regardless of who wrote it, it could still be poor, irrelevant or misleading.
I'm not defending AI over-usage, but you folks evidently haven't thought about any of this nearly sufficiently enough... and the policy doesn't seem to be addressing what it ought to be: whether everyone is actually thinking, learning, discerning, deciding etc...
Ps I passed your comments to chatgpt and asked what it thought about it all, and it missed all of this. But then agreed once I brought it all up. Then it did a poor job of writing it's own response. I couldn't be bothered to guide it towards something more appropriate. https://chatgpt.com/share/6a86cc31-f058-83e8-8abd-3921b518ec...
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
I think your linked chat demonstrates quite well why you are wrong.
The sender's burden is a few lines of text "Write me $FOO, conclude with $BAR".
The recipients burden is an order of magnitude greater, trying to find your conclusion in hundreds of lines of text.
After all, your initial line:
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
Is concise and clearly understandable in 21 words . I understand your point even if I don't agree with it.
The linked chat has 1424 words. Your point is in there somewhere, but now you want the recipient to play "Where's Waldo" to find your point.
(BTW: I can't believe just how effectively you argued against your own point! I have never seen someone do it as effectively as this in all the time I have been reading HN)
There's so much wrong with your comment that I don't know where to start.
You say that my initial line is good, but then that you disagreed with it without saying why/making a counter argument.
As for sending anyone on a where's Waldo adventure, you've fundamentally missed the point, which was completely self-contained in my comment - the postscript link to chatgpt was just, at best, reference material for anyone who cared to look.
As for the fact that chatgpt did a terrible job of it all, that doesn't undermine my point even slightly. Because my point, made very clearly in my post, was that you can't just offload everything to AI. You have to, at the very least, give it context, direction etc. At that point, it can be quite useful. I use it effectively all the time to polish my writing, remove my worst tendencies etc.
If I cared enough, I'd have had it polish this response to make it significantly less aggressive while making the same point. But you don't deserve that, because you didn't put in even the slightest effort to communicate in good faith
>But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
Not the same burden.
The value of the person is digesting and giving the materially important stuff in the answer (including their expertise and opinion). Not the AI diarrhea.
I made some ninja edits, so perhaps you didn't see them. But it is entirely possible that both a person and an Ai could produce writing that is both good or bad, useful or misleading.
You can also direct the Ai to write about specific points and, again, it may write somehing good or bad. I provided a link to a chat I had with chatgpt about this very conversation and it shows everything I was saying.
The fundamental issue is whether people are taking responsibility to communicate what needs to be communicated, in an effective manner. Whether the final words were authored by Ai or not is largely irrelevant.
>I made some ninja edits, so perhaps you didn't see them. But it is entirely possible that both a person and an Ai could produce writing that is both good or bad, useful or misleading.
Theoritically yes. In practice:
- AI responses I get are walls of written diarrhea (even when they're relevant and take our particular context into account). The one's that don't are from people who write well based on their expertise, and who know to edit AI responses and not just paste them.
- I'd rather get the bad/misleading/not-useful response from the person, rather that the good one they made via AI. This way I know they don't know what they're doing, are bad communicators, depend purely on the AI, and thus their position can be eliminated or replaced by someone cheaper merely relaying to/from the AI.
> - AI responses I get are walls of written diarrhea
Evidently the people you are communicating with are not good writers/thinkers. That's the prerequisite for effectively using AI for writing help. Just as its the prerequisite for using it to produce code. If you can't code, you can't get Ai to produce good code.
So the real issue here, perhaps, is "vibe writing" rather than "Ai assisted writing".
> I'd rather get the bad/misleading/not-useful response from the person, rather that the good one they made via AI. This way I know they don't know what they're doing, are bad communicators, depend purely on the AI, and thus their position can be eliminated or replaced by someone cheaper merely relaying to/from the AI.
Agreed. But if people who don't understand the issue are feigning competence via AI then that's a fundamental problem of not just competence across many domains, but also character. As I said in another comment, the policy should be something to the effect of "you're paid for competence, integrity etc. If you don't have that, we don't need you. Take that into account when you use Ai"
> Theoritically yes. In practice:
I get the feeling this is the whole disconnect. People have high expectations from AI but haven't really been on the receiving end of a massive amount of text/code generated by it.
We recently had someone fired for using AI excessively. Before the firing, whenever anyone on the team sat down to give feedback, the only answer was "but everyone is using AI, we are supposed to use AI, how can it be bad?".
Until their boss sat down and forced them to read the verbal diarrhea and they finally admitted that it was 90% hallucinations and AI-filling-the-blanks.
Yes, so the issue here was not AI but a lack of competence, prudence, integrity, responsibility, discernment etc.
It's the equivalent of vibe coding vs Ai assisted coding.
Vibe writing vs AI assisted writing.
The thing is, I expect the person who sent me something to know about and understood what they have sent me.
I don’t expect the other person to copy my response to the same chat and paste the AI response back even without reading it.
Don’t be a meat proxy.
Well, In any healthy organization or relationship, there should be no problem with someone not knowing about/understanding something.
If someone is feigning competence, that's the fundamental problem, not AI
> The fundamental issue is whether people are taking responsibility to communicate what needs to be communicated, in an effective manner.
While I agree with your overall distinction, I believe the scenario you portray is so unrealistic that you will never encounter it.
Yes; if I could wrangle Claude into communicating exactly as I would have, the practical difference is of course not observable. But that's never happened, and I've had an enterprise license at work for years.
The value in having a human write it is that whenever something bad/misleading/redundant is introduced, that person will learn from the feedback, but an AI will at best have some memory/skill/prompt nudge for next time that might have a shallow/localized band-aid effect, but will never stop doing similar mistakes elsewhere.
> The value in having a human write it is that whenever something bad/misleading/redundant is introduced, that person will learn from the feedback,
For me, the value of receiving comms from a human is different: it communicates not just the idea itself, but the fact that the human already understands the idea.
The people arguing that round-tripping their thoughts through AI still means that it's their thoughts apparently also believe that the game of telephone always results in perfect comms.
It's not just AI; adding more hops to a route changes the message when each hop is a human too, it's just that we notice it more when the hop is an LLM because of just how different they write compared to humans.
> if I could wrangle Claude into communicating exactly as I would have, the practical difference is of course not observable. But that's never happened, and I've had an enterprise license at work for years.
Sorry to say, but this is a skill issue (and not necessarily with using Ai). I've used Ai MANY times to help SIGNIFICANTLY improve things that I've written - from a few sentences to many paragraphs.
As I've said in other comments, it'll rapidly and effectively remove all my worst tendencies. And, likewise, sometimes I'll have it write a first draft when I give it appropriate context/guidance that I can't be bothered to fully curate/synthesize. Then, once it gets that draft, I'll take over and tidy it up. You might say that the "can't be bothered" part is the entire issue, and I'll again say that what matters is the final result - a master chef or chief architect is not doing all the grunt work...
> Sorry to say, but this is a skill issue (and not necessarily with using Ai). I've used Ai MANY times to help SIGNIFICANTLY improve things that I've written - from a few sentences to many paragraphs. As I've said in other comments, it'll rapidly and effectively remove all my worst tendencies. And, likewise, sometimes I'll have it write a first draft when I give it appropriate context/guidance that I can't be bothered to fully curate/synthesize. Then, once it gets that draft, I'll take over and tidy it up.
Isn't this contradictory? Why do you need to take over if it's doing such a good job?
For me, being quite "dyslectic" - LLM generated sentences have the advantage of being spelled correctly and being grammartically correct. I rather prefer iterating over LLM text than writing it by myself. Example: Typing this simple text above by myself was not straight forward.
Write yourself and tell the AI: "Fix the spelling and grammar. Do not change the structure or meaning" then.
My problem isn't who wrote it, but the pretense of something having written by a person while it wasn't.
There's something profoundly disrespectful about it towards the receiver.
Oh very much agreed. It fundamentally destroys someone's reputation when they refuse to acknowledge that Ai wrote something.
Far better, and frankly perfectly acceptable, to say or have it implied or concluded, that Ai helped with the writing in some way. Still, if you're paying attention, it often shouldnt really be all that detectable because you should be aware of AI's smells and you can edit them away. Though, sometimes I'll leave it as obvious that Ai produced something I shared - eg making a well formatted github issue or whatever. But I'll also make sure to convey that a thinking human was behind it, vetted it etc
There's something profoundly disrespectful about feeding snippets produced by human authors en masse into a machine learning algorithm to create a predictive text generator to be applied as labor replacement, regardless of accuracy or appropriateness to the context or use case. There's something profoundly disrespectful about forcibly scraping the human corpora without asking for permission first, or considering compensation in anyway, or just setting out that "yeah, we're doing this, and damn the consequences". There's something profoundly disrespectful about then taking this device running on investor money chasing the promise of a value that just can't conceivably be delivered, and employing it to do all the unethical things you could imagine just because you can engineer the "No" out of the model.
Of course the level of moral disengagement that's rife around here generally precludes any recognition of such things, as that might be construed as "harshing the vibe on the innovation, man". So I guess everyone had better just settle in and get used to some serious profound disrespect until this type of thing hits enough of a critical mass that a societal ethics framework begins to manifest around it and not be ignored.
> But isn't that "burden" on others always going to be there, whether you or AI wrote it? That's what communicating is...
I think it can be OK if AI is used to polish up one’s message or to help structure it. Basically help with grammar, formatting, and wording.
But copy pasting an AIs response to a question? Not at all. People earn a reputation over time. TRUST is earned. An AI cannot earn trust. If you copy paste an AI response verbatim, i need to read everything in your message with a big grain of salt. Imagine doing this all the time as a reader? Exhausting and a waste of others time. If i want to read raw AI output i’ll ask the AI directly, thanks.
If you write something in your own words, or at least synthesize an AI response before sharing it with me, then i know i’m reading something that has at least passed the smell test.
It’s the difference between reviewing a pull request from an intern and a principal software engineer.
I don't disagree at all. But it still remains that AI can very much be effectively used either to write first drafts OR refine your own first drafts.
Though, I think such a writing partner can only really be useful if you know how to write yourself, which also requires knowing how to think (because writing is really just thinking).
So, again, in the end all that really matters is whether you are effectively communicating what needs to be communicated, Ai-assisted or not. It's quite apparent to other thinking people when someone is not doing that, and that hurts their reputation, undermines trust and ultimately leads to losing your job, connections etc...
So, again, the policy really should just be something to the effect of "you're being paid to help us be effective and correct. If you can't do that, we don't need you. Take that into consideration with regards to the tools that you use.". That's obviously not an adequate policy - Ai could surely be helpful in polishing it! But it most definitely couldn't get to the root of the matter on its own.
> But it still remains that AI can very much be effectively used either to write first drafts OR refine your own first drafts.
There is a world of difference between this and copy-pasting LLM output into slack and email.
You can write the first draft with the LLM, or have the LLM help you do an analysis, look at how code works, and so on. But you need to check for the LLM to be correct and you need to be able to articulate what the LLM found in your own words. If you don't you have just offloaded all the work to someone else: someone else will have to check, verify, investigate what could very well be an hallucination (bad context, random bit flips, whatever the cause doesn't matter). Your effective contribution to the discussion will be negative.
You bring up a good point. Maybe it doesn't matter whether an ai or a person wrote it, as long as it is concise.
And because most ai's are not concise we blame ai's for faults we don't like in humans either: being longwinding
Well, I'd say concision is not the key - sometimes you should be verbose or something else.
In fact, my tendency is to be far too long-winded, with lots of perhaps-unnecessary parantheticals etc. Ai is often shockingly useful at reining me in and focusing on the issue at hand, as well as removing my worst tendencies towards meanness etc...
Im a good writer, when I want to be. But having Ai review, edit, write drafts etc undoubtedly makes me a better writer/communicator. But only because I'm actively thinking about it all, rather than just saying "respond to this" and then providing copy pasta.
You're missing the whole entire point.
Good luck with that, I hope you are in a better org that I am in that regard. I wanted to do that, proposed it in a directors' meeting and got push back from directors already using the Slack MCP to post normal conversations on their behalf (90% of time, pure slop).
What you wrote just seems common sense to me, and I'm hopeful it will emerge as such given enough time.
Thanks for sharing your thoughts. One of the best frameworks I have encountered, and one I point people to whenever I can, is the AI Fluency framework by Anthropic, Prof. Joseph Feller (University College Cork), and Prof. Rick Dakan (Ringling College):
https://www.anthropic.com/learn/claude-for-you
This framework provides a more abstract way of thinking when working with AI systems. Especially valuable to me is the pattern of the 4 D's: Delegation, Description, Discernment, and Diligence, which I have now internalized whenever I use AI systems. It helps me understand my own and others' (mis)use of AI systems a bit better.
Nice. This is exactly how I mostly use AI. Doesn't matter if it's code or an e-mail. I have the AI do the tedious work of ... well, uh, "working", and then ideally learn from it and then transform it into my own work, as I would have done it without AI, but just being much faster now. It also establishes a natural threshold of time/work ratio for me that keeps me from using AI for every little thing. Also it feels like I am learning something new with every prompt/response.
Any chance you would publish that guide and share here?
Seconding this request... Our team is going through AI adoption without a clear direction and we're getting to the point we need to sit and come up with some guidelines. It would be awesome to have some human-made examples as inspiration
Why write all that when just: we do not use AI in client or intercompany communications?
Like is your insight that copying and pasting something is not understanding it? Do people seriously not known this?
There is no need to Socrates when writing policy docs. Maybe this is why people ignore reading your actual voice messages that you feel a need to also justify this basic request. Sometimes the human voice thinks too highly of itself also.
Because if you think and act as your colleagues are just disposable soldiers that must follow direct, simple orders, then your org will eventually be made of that kind of people.
And maybe GP doesn't want that.
> You’re being paid for your expertise and insights.
For purposes of training the AI (enablement firm)?
It's sad that there is a chance people put your guide into AI and ask it to "summarize"
They will paste it to their agent and ask it to live up to those requirements
Sharp idea. Thanks.
Am I the only one suspicious of your comment?
> Write as yourself.
Unless it's your first day on HN (it's not 2008 account), most people use paragraphs and sentences longer than 2-5 words. Something like "You, directly." in the middle of all that has a certain vibe to it.
> Copying and pasting Claude responses... directly shifts the burden of comprehension
Most definitely.
The problem is we live in a world where people try to pass off copy pastes as their own all the time, many even smart enough to remove the em-dashes now. So I don't think that's going away soon. Your first sentence seems human, the rest doesn't match the tone at all.
This is the sharpest point. You genuinely delved into the commenter's prose and found the goblins. It's not just the ground truth–it's a load bearing, belt-and-suspenders approach.
>Most definitely.
Isn't that the very thing you just claimed people don't do?
For simple, impactful sentences, I often use short ones. They work better.
I’m not convinced tbh, a lot of my colleagues messages are “x broken” or “I do y” … no context whatsoever, and then I have to coax their context for making these decisions.
Since people started using Claude now I get full context of everything… might be too much sure, but to be honest over-communicating seems better than under-communicating - sure it’s boring and tedious but that shifts the blockage to me.
Otherwise each of these coaxing sessions is something I have to keep in my head until resolved, which is a load in and off itself.
And even better - I can point Claude to that message and gives me a summary. You might say this is silly because we are paying the LLM tax, but knowing what to share, and then verifying against my situation are two different things.
Sometimes they have checked the wrong thing, sometimes they need guidance, sometimes I need to investigate before I can answer. A full LLM message with context is like “summary of their working context that I can resume on my end” so I don’t waste cycles asking or rechecking etc.
And it is still possible to be a learning experience for both of us since the initial message is the start, we can then talk to each other like humans, both much more in sync than before…
It is kinda ironic but I think it does push things in a better direction. Would I have loved it if the message was human, direct and to the point - obviously, but we are all busy, we got shit to do, and this is a useful “resume” mechanic
> Since people started using Claude now I get full context of everything…
And I think this is where one key argument on the page itself I think fails:
> The person on the other side has the same tools you do.
Yes, the person on the other side has Claude; but in many cases, their Claude doesn't have the same context as my Claude. Saying, "This thing happened in an area of the code I don't have any expertise in. Below is Claude's analysis, which I'm not qualified to assess" is I think a valid time to copy and paste LLM output. [1]
The other thing to say is that it can be a "LMGTFY" style response; i.e., "Why did you ask me that, when you could have gotten a perfectly good answer from ChatGPT?"
[1] In the particular case I'm thinking about, it was a bug in an ocaml component; I don't know ocaml or need to know it; handing Claude's analysis 'caveat emptor' to the maintianers who do know ocaml I think is more helpful than not.
> Below is Claude's analysis, which I'm not qualified to assess
I’m reading a bunch of game modding discords, and this exactly what happens, and what’s hated. Because Claude is also not qualified to assess it. Not by default. The modder’s Claude might be because of the extra context and steering, but the user’s produces usless slop, instead the actual log, and context what the user did is what’s helpful.
I’d say this (when you don’t understand it yourself) is pretty much the worst time to send AI output.
Quick edit: Yeah, I get it in situations like GP where the user is unhelpful, doesn’t want to learn, doesn’t care, and you also can’t tell them to fuck off. But just because the user gives Claude the context they should have given you in the first place, doesn’t make it good.
If they user types "x broken" or "i do y" into Claude or ChatGPT with no context, that's absolutely no better than just sending it to you.
I'd argue that it's worse because now you have no idea what the actual original context was, just the AI's hallucination of it.
You'd think two decades of StackOverflow would've teached people how to ask a good question.
LLMs have an equalizing effect.
> a lot of my colleagues messages are “x broken” or “I do y”
These people had below than average communication skills (or at least they were putting below than average effort communicating with you), so LLMs improved their output.
However, the society kinda expects people can and Should improve themselves over time. The issue is that LLMs improved their output, but not necessarily them.
I take your point but I have two counterpoints:
1. I've experienced both (ai paste slop & ai providing helpful context that would otherwise have been absent) but the former has been a much bigger problem than the purported benefits of the latter
2. before the advent of LLMs, I had fairly strong contrarian opinions on the latter (clueless users asking questions with no detail) that I still hold to this day. A good proxy for my opinions on this is that I think https://nohello.net/ is a crutch for undersocialised (or overworked/near-burnout OR undiagnosed autistic) engineers & should not be normalised. The origin of this sentiment is users entering a quiet IRC room & saying "hello", then leaving & not being around when someone eventually replies: in that context nohello is pragmatic, but taking it beyond that to modern high-frequency realtime chat is an antisocial overstep. Leading with the entire context of your query may have some practical benefits but they're honestly overstated in exchange for killing natural social tendencies towards human dialogue. Similarly, a non-technical user giving you zero info about a problem is an opportunity for conversation & mutual education. It's relationship-building & helps build a foundation for further interactions with this person. It was never a problem in the first place; "solving" it is just another milestone on the road toward de-socialising humanity.
[delayed]
You are right - and this has been the case for me before, but unfortunately human resource constraints are not something most orgs optimise for, including my current one.
I would love to spend more time and talk to the people with their issues so we can come to a common understanding… but neither have the capacity for this, and now things are getting worse I think.
Maybe someday agents are good enough that we just discuss direction and implementation details, and every “developer” becomes more of an executive in a company, free to chat about more than just work, focusing on building human relationships and leaving “the tech stuff” to their agent teams…
And old geezers like us who actually know how to write ourself out of a for loop would be the premium - wizards holding forgotten knowledge, incomprehensible to mere uninitiated mortals. I wonder what distinctive dress and hair styles will we develop :-D
Points to the future where my bots talk to your bots. I think that’s healthy.
Context is everything.
Btw, Claude code sent my first accidental AI email yesterday. I guess permissions changed? Before it could only draft. Uhoh, waiting for the onslop.
Hah, I guess we are not far away from the “Avogadro Corp” future - and that book was written just _before_ LLM, chatgpt and the attention is all you need paper :) That author had quite the insight
If you're absolving the other person from any understanding of the issue and the alternative is "stuff broken" then yes what you say makes sense. But that's an incredibly low bar and one would have to say - that's not a colleague, that's an end user type relationship.
> I’m not convinced tbh, a lot of my colleagues messages are “x broken” or “I do y” … no context whatsoever, and then I have to coax their context for making these decisions.
> Since people started using Claude now I get full context of everything… might be too much sure, but to be honest over-communicating seems better than under-communicating - sure it’s boring and tedious but that shifts the blockage to me.
Yes, but the kind of engineer that used to say "X is broken" will send you some slop with missing context and Claude will try to apply general solutions from its training data rather than from the specific context. As somebody on the Ops side that usually helps devs fixing things here and there, in those cases I'd rather them say "X is broken" and then I ask the question to Claude with all the context I have.
Probably the best solution currently is a shared skill with at least some curated context.
> sure it’s boring and tedious but that shifts the blockage to me
I think that's the key. Generally people would rather not receive slop because the receiver has no idea how much thought the sender invested in the communication.
In your example, the more context the better. I think there's an interesting nuance there.
> the receiver has no idea how much thought the sender invested in the communication.
That's an interesting thought. The message can have metadata that says the sender's spent 1/5/10/60/120 minutes or some other metric of how much thinking over the communication happened. Just need a way to monitor people's thoughts and digitize them to make that happen.
> Since people started using Claude now I get full context of everything…
Really? How does Claude gather that context?
yes, its all ai writing. A bunch of the metaphors here dont make sense.
"read it all, yes even the bullets. especially the bullets"
if you read one thing from a large wall of text, its gonna be the bullet points, right? thats a big part of what they're for?
"It is the conversational equivalent of forwarding the email."
this is a terrible metaphor. it seems to rely on the idea you would be mad if you sent an email to someone and they forwarded to someone else? but thats useful, theyre literally sending it to someone who is presumably better placed to help you, or raising it to someone more important because they think what youre saying is important
Interesting to note the page also ends with "written by a human, on purpose"
Yeah, it's obvious enough on its own, but I'll add that Pangram also thinks it's 100% fully AI generated.
Hypocritical to say the least, given the message.
Pangram sucks
I've never seen a false positive from Pangram for human-written text of 300+ words.
Pangram sucks by seeding fear.
The angry version is even more obviously AI
Deeply ironic that it sure looks like taking the original post, pasting it to an AI with the message "make this angrier" and then simply copy-pasting the output back to the website. In direct violation of their own (supposed) principles.
And the result isn't even very angry at all. It's like mildly peeved internal company communication (also in direct violation of "don't send this to your manager")
I actually wrote it all myself... The original version was the angrier one only, but people complained that it was too angry, so I tried to smooth it out with another version people could use.
I had the same feeling. The terrible metaphors smells like Claude.
> especially the bullets
DS9 Garak "especially the lies" reference maybe?
This was actually a small rant from that same day, a colleague had pasted a list of 20 bulletpoints, all of which had some weird stuff in it but overall they were right. This is why I added "Especially the bullets", English is not my main language so I tried to translate myself, but vocabulary sometimes is lacking, sorry about that
Same but shorter: https://noslopgrenade.com/
I remember back in the day when you could apply this to Google search results. When I was freelancing, I had a number of people I worked with who would fire off fast questions, but they would still pay me for my time. Then I would return what was essentially a Google search result shortly after. Yes, I would send my own answer, not some paste from Google.
But the point is. They had search, so why didn't they think to do the search? They had the same tools as I did. And yet regularly they couldn't think of the answer. I still deal with this with AI today. When you ask a question, maybe you're asking for the interpretation of the question by the person on the other side. And it's that interpretation that goes to the LLM. Or otherwise some filter of how that person thinks about the question differently than you do.
Translating a client question to an actual technical question is the hard part I guess. Google-fu is still a thing, and absolutely a thing you were/are getting paid for it seems.
> It's a great drafting partner
This I have been doing a bit. I strictly asks it to never give me suggestions but raise questions, concerns, tell me if it feels if something can be better, style-wise or otherwise without giving me those options.
And a bit of other overall feedback requests.
And if the model adheres to this and doesn't jump the gun, it indeed works quite nice. But it often jumps the gun, forgets, doesn't give a f- about my instructions, or whatever. Maybe more robust local models will better.
I have a colleague who just paste PR reviews unedited and unread into git. It has been driving people insane and, when people do inplement what it say, straight up degrading the codebase.
This is exactly what I have been feeling in response to this person. They are not contributing. They are creating work for others by pasting every possible problem they can find, most of which are out of scope or context, and letting the requester sort through them which takes hours if you take it seriously.
Some people do and implement everything he asks. They are going way outside the scope of their ticket and fixing every problem in every file they touch. It is chaotic and not good for the codebase because a lot of the problems are repo wide and should not be fixed in isolation.
He doesn't seem to understand why this might bother people or why it takes them so long to respond. He suggested just pulling the comment into Cursor and just asking Cursor to respond...
At that point we are just straight up vibe coding with no human in the loop.
That being said, Cursor's bug bot does a way better job reviewing, and people just ignore it when it asks for something out of scope because it's not a colleague claiming their code is broken.
Luckily I don't have this problem since I am so full of shit that one of the things I still enjoy the most is answering people with my own crap.
Ironically, almost all the content on that page sounds AI generated to me!
Yup, I think the author needs to take their own advice and invest some time into putting it into their own words.
Yup, that's the biggest joke, the website is not worth the weight.
Layout, colors, font and especially language all immediately scream AI. It's fascinating how many of these articles make it to the front page. I would think almost everyone here would be an expert at spotting AI generated slop by now, but people keep up-voting it.
just me or does this entire site/text read as 100% LLM generated? extremely ironic if so, they just pasted the AI.
I got that impression too. Absolutely reeks of LLM. Really irritating.
DontMakeWebsitesFullOfObviousAISlopThatComplainAboutUsingAITheIronyIsTooThickForMe.com
same but shorter (already posted on hn): https://gruhn.me/blog/2026-08-03/
Dont be a meat proxy! simple
Actually tbh AI outputs can be pretty good and I have been surprised to see a lack of copy / pastes of them in the format of:
Unique human prompt + AI output which is useful and does a good job expanding on the prompt topic and which when pasted saves people the time of having to go prompt AI on the same topic (fully reviewed by a person before sharing it, to ensure the content has relevance)
Agree that it isn’t nice to just copy & paste, but the assumption that your AI generated answers will the same of someone’s else is wrong. Everyone using AI for a while built a different set of preferences, memories and context in general which influence the answers.
Or no memories or history.
Would be interesting if someone said “I spent $2.50 using llm-council with the three top SOTA models and think this reply makes sense: {paste}”
(But better “not sure, though after spending a few bucks on model APIs, perhaps {human-written analysis} is in the right direction”)
Also like you, I did think on the angry version “they have the same AI you do” isn’t likely to be right in the consumer space since their AI is ChatGPT (free version?!) and yours is Claude or something.
You're giving too much credit to people. The folks who run an LLM council aren't the ones blindly copy pasting things.
There's responsible AI use (which I think you - and I, and many people here) do, and the "I don't care, I've outsourced my brain to OpenAI" type of people.
Even with the same model results are different. I work in a startup where they went bonks with AI ~1 year ago. Every employee got a Claude account that I don’t even know what tier is (we just have what feels like unlimited tokens), they implemented infamous AI usage leaderboards, people were getting “you’re not using AI enough” feedback and all of that…
Over time, it was impressive to see how much the output of the same prompt would differ between us, and it seems to be heavily influenced by the role of the user. I.e. answers coming from the agent of a Data Engineer, a Backend Engineer and a Domain Researcher would not be the same at all. Some of us use harnessing tools, some not, but regardless it we saw differences.
So, for example, when a backend engineer would send a question to a data engineer, there was *some* added value even when it was a lazy copy & paste.
Similar thoughts here. My gut tells me "if I wanted AI responses I would just ask AI", but in truth LLMs are not like Google Search. LLMs are more like a slot machine where you put your prompt in and out comes a combination of insights that are personalized to the inputs.
The premise of the article is that everyone has access to the “same tools” but that’s glossing over the specialization inherent in a specific expert’s harness config (memories, skills, rules, repos, etc) that produce actually helpful responses (eg engineer’s view vs product manager’s view).
The article says they could have asked an AI themselves, but that is missing that I may have an agentic context with exactly the right information to answer the question well, which they do not have. For example because I worked on something closely related. Such an agent output is very useful, and taking the time to rewrite that to a short personal answer is not always worth it.
Your agentic context is still going to shit out 3 pages where a sentence would suffice. Use AI all you want, and then WRITE THE ANSWER IN YOUR OWN WORDS.
> Don't Paste the AI, please
You're gonna need a stronger language. Why do you think people love AI? If I have to think and god forbid read the response I might as well live in the woods or something
Linked a the bottom of the post is the angry version https://dontpastetheai.com/angry/
Sorry this website is too polite.
DO NOT FUCKING PASTE AI RESPONSES TO ME, EVER, PERIOD.
When someone tells me they "talked to Claude" they sound retarded. Just tell me you looked into it, and I fucking swear to god if you just took what Claude said verbatim and do not understand the problem then you are a cunt and I will stop working with you.
When someone pastes LLM output in chat I completely ignore it. I will just continue the conversation as if they said nothing and silently mark them as cretinous in my mind.
This is not up for debate, you will talk to me in short well formed sentences and get to the fucking point or you will fuck off
I like this comment so much I checked out your profile. Turns out I just like the cut of your jib.
Very ironic this page is full of AI text
Unless I'm your research assistant, why would I be asking AI questions for you when you could do it yourself? If you need an answer I already have I will gladly share it with you. If you need my help to find an answer, then tell me that and I gladly will help.
Thanks for sharing my little rant again <3 we are looking for translations too!
I don’t think it’s that Sol loves to cheat; rather, every GPT model since 5.4 seems to have had that tendency.
Perhaps you meant to reply to a different post?
What if i have used the AI, it answered, but I don't understand the answer?
If you don't understand the answer, don't answer somebody else with it. You're attaching your name to whatever you copy-paste to people, and if you don't understand it there's a chance it's complete nonsense. Save yourself the embarrassment. If you don't know how to answer somebody's question, just say "I don't know".
Give them a second brain so they don't bother you with silly questions: https://github.com/amaldini/maldalang/blob/main/Examples/Age...
What if English is not your first language and you are afraid of sounding silly when you email or text someone?
Don't be. Read it through a few times and fix anything you think needs fixing. Go simpler if needed. You might just use shorter sentences. They are fine. And then you can use some pre-AI tool like Word that has some suggestions.
Ask the ai to point out the egregious mistakes only.
Especially today, I will prefer your slightly funny English to something that feels like AI.
And personally I do Support for people around the world and I often ask people to write in their native language.
Then I can translate myself and if the default ai translation gets something wrong I can add necessary context to improve it - which I can't really do if the only thing I get is the bad ai translation.
If you respond to me with slightly "off" english, I at least know it was a fellow human who wrote it and I will act accordingly.
How will you learn if you don't practice?
You write your reply and ask the AI to fix the spelling / grammar for you. It's really not the same thing
Create new chat after step 2, then ask AI to fix the spelling / grammar of those three sentences
How did people do it before AI? Spellchecks have existed forever. Also, people have the ability to improve, so if you offload all your work elsewhere, you will never improve.
خب، در آن صورت، من وقت می گذاشتم و متنی را با دست می نوشتم و دقیقاً همان چیزی را که نوشته بودم از طریق نرم افزار ترجمه ماشینی اجرا می کردم.
Да, тоа изгледа како среќна средина.
det sker nogle gange! når man er ny til et sprog, det kan være forskrakkelig at skrive i et professionelt område. Men det er bare en del af dine udvikling - du kan sagtens skrive på hvilken sprog når som helst, og de kan bruge deres værktøj til at forstå det, du siger.
Der skal kom et tid hvor du skal tale på en sprog på dit kontor. Uden hjælp. Du har bruge for at øve dig, og folk er alligevel pænere end du tror.
(this being written in a foreign language I am learning is part of the point)
At this point I would hold that silly sounding message in a much higher regard than I would hold AI slop.
I also suffer from this! But it's called growth. You will not improve language skills if you are too afraid to practice them
And in my experience, us Brits, and USA folk, we tend to be quite understanding of broken English
Use it to rewrite your OWN thoughts, or translate it to english. Not bloat with phrases sounding fancy but carrying no meaning
This does feel like responding to rudeness with more rudeness. My personal guess would be that people who do this (pasting AI responses) know what they are doing, not genuinely trying to help. It's the equivalent of sending LMGTFY.
Disagree - most people aren't bothered to ask AI and have a conversation with it, they will go to you to get an answer so they have someone to blame when it's wrong.
I don't really seem to be sent very much unfiltered AI output - do other people? I'd love to hear horror stories.
This is like an earlier internet era when it was common for people to just copy-paste directly from other websites into a forum response.
I think this text also applies well to code written using an LLM.
"written by a human, on purpose."
idk about that
Worst case you just ignore it. Best case there is something useful in it. Shrug.
The inverse of this is. Don't ask me a question a 9b llm could have answered.
If you are getting blocks of llm text perhaps it's because you don't deserve their input and could have satisfied your query by not making it someone else's problem.
I don't understand the need of need "Satire, mostly." at the bottom, are opinion on things so dangerous?
I added that because people started attacking me since it was a bit angrier at first (the angry version was the only one) and they didn't want to send to colleagues or anyone else, just close friends so they asked me to add a smoother version and to say it's a joke
People are doing this? Wow, I'm sorry folks.
I was recently listening to a podcast about mentalization theory therapy methods for BPD .. for a friend .. Mentalization being the activity of understanding whats going on in other peoples minds, or in ones own. And they idea of pseudomentalizing or "pretend mode" was put forward, where a person looks or sounds like they are .. engaged in mentalizing communication, but are actually checked out. Then I started to think about all the AI slop going around .. and the phenomenon, this meme site is responding to, or AI posting responses.. and I wondered, is this a symptom of.. a global society wide phycological functional deficit, where A) we're none of us communicating as if there is people on the other end.. and B) we're not thinking about our own minds, and taking time to understand things..
[edit] to extend the connections The charm of a messaging board like this is bringing "join attention" to a topic, not just sharing separate monologs.
And also, always wondering if something was written by a human or statistically generated could be seen as Epistemic Hypervigilance, leading to stress, and loss of epistemic trust. This could reduces opportunities for taking on new information of feedback, and cause rigidity and isolation.
A much better and straight to the point example is https://noslopgrenade.com
Don't be a MEAT PROXY
https://dontpastetheai.com/angry/
Almost certainly written by AI
... the page never argues that AI is bad or not useful broadly. Not sure if this getting commented again and again is because people are just that bad at understanding that context matters. Not everything has to be evil / slop or pure / virtuous, regardless of context.
no, but it does argue specifically against presenting people with large amounts of ai written text without reading it or checking that it makes sense.
It also ends with "written by a human, on purpose"
No. Don't ask a question that AI can answer.
Similar in spirit: https://stopsloppypasta.ai/en/, discussed a few months ago [1].
[1] https://news.ycombinator.com/item?id=47389570
Pasting a blurb of AI is sometimes a form of "Let Me Google That For You" LMGTFY. If someone asks me a dumb question, I will often resort to giving them an AI answer, that they could have easily looked up themselves.
It literally takes you more effort to get an LLM answer, and if it doesn't, it's not really a dumb question, is it?
> If you don't have anything to add, it's okay to say so. "No strong opinion here" is a real, helpful reply.
This is really useful, I wish more people would understand this: you don't have to have an opinion on everything. In fact, you probably shouldn't have an opinion on most things.
My own take on the AI issue: https://blog.gpkb.org/posts/just-send-me-the-prompt/
This resonates. Had this in the start of the year with clients sending me their “research”, which, of course, was just a Deep Research result. Lead me to send this article when I work with folks: https://www.marcia.no/words/etiquette
I like to think it helped, but likely what helped was that nobody seems to use Deep Research anymore, haha.
Ironic for this to be full AI Slop.
"Angry" version is clearly 1-to-1 ai output. I think I prefer dontbeameatproxy
that's the joke, isn't it?
They did not read AI output, they will not read that page either.
It's the modern era lmgtfy
personally i like editing much more than writing (the scare of the white page or how it was called…)
and i detest ai slop
so my solution is to speak combined with an ai cleanup like macwhisper or wispr flow of all the uhs and ahs and repetitions or rephrasings etc
you can even use ai to pretend a conversation or interview style
then i have stuff to edit
boom
I can really only imagine the social context you would want to use this response in would be inappropriate / anti-social. If you had a well-meaning work colleague send you something AI, sending this back is kind of not a good look. If it's meant to be an education piece we got the memo...
Weirdly, this post speaks for everyone.
This site is AI slop.
Did they really just say "genuinely useful" in a post decrying the blind copy-pasting of AI text? (it's a strong Claude indicator)
lmao love it. Although I'm guilty too though
> A short reply from you beats a long one from a model, every time.
Disagreed. Bad writing from a chatbot beats worse writing from its user every time - regardless of length.
One reason we have so much bot slop is the attractiveness of bad-writing bots to the many people whose writing is even worse.
Okay, boomer.
Beautiful theme. Author has style.
Slop style.
Bollocks, the other person DOES NOT have the same tools as you. They may, but to assume they do is incorrect.