AI and Gentoo forums posting

[pingtoo]

I am happy to clarify anything I've said, but when you put my words into LLM ask it to explain them I see it as a little rude. I don't know how much context you have given the LLM - have you fed it the whole thread? Does it know why I said the things I did? Can it? The person I am facing is no longer you, and I quickly lose my patience if I think I'm talking through someone to a machine. I am not judging you by the words the LLM has written, but it takes a lot of effort not to do that, because we usually do judge other people by the things that they say.

Sincerely, there is no grudge. I hope we can both work together on this.

I am happy we can find a way to work together.

So following is exactly what I sent to Gemini

Please analyze the tone of following sentenace, does it somewhere contain hostile attitue?
"Why do this, though? Either you can speak with experience or authority on a topic (and help other people with it), or you can't. Do you think people asking for help want someone to come along and say "well, I don't know how to do the thing you need, so I plugged your query into an AI chatbot and juggled the words around a bit"? Wouldn't they just do that themselves? It'd be better to get an AI answer you asked for than an AI answer in disguise"

I have explained earlier I am not sure but I do suspecting these sentences hostile intention, may be I could have ask my daughter but she is not available at the moment and I am very curious does Gemini have concept of human intention. so I post it in to Gemini and the answer is

The tone of this passage is **confrontational, skeptical, and critical**. While it may not reach the level of "hostile" in a threatening sense, it certainly carries a high degree of **passive-aggressive friction** and intellectual condescension.

Here is a breakdown of why this sentence feels aggressive:

### 1. The Use of Rhetorical Questions

The author isn't actually looking for an answer when they ask, *"Why do this, though?"* or *"Wouldn't they just do that themselves?"* These are "loaded" questions designed to make the recipient's actions seem illogical or foolish. This is a common hallmark of a hostile or dismissive tone.

### 2. False Binary (Either/Or Fallacy)

The statement, *"Either you can speak with experience... or you can't,"* creates a rigid ultimatum. It leaves no room for nuance, such as someone using AI to help synthesize complex information they *do* understand. By framing the issue this way, the author is essentially telling the recipient they are being "useless" if they aren't an absolute expert.

### 3. Dismissive Language

The phrase **"juggled the words around a bit"** is intentionally belittling. It reduces the effort of the other person to something messy and thoughtless. Similarly, calling the result an **"AI answer in disguise"** implies that the recipient is being deceptive or dishonest.

### 4. Sarcasm

The imagined quote (*"well, I don't know how to do the thing..."*) is a form of strawman sarcasm. The author is putting words into the recipient's mouth to make them sound incompetent.

---

### Summary of Attitude

| Trait | Presence | Explanation |
| --- | --- | --- |
| **Direct Hostility** | Low | There are no insults or threats. |
| **Condescension** | **High** | The tone suggests the author is "above" the behavior they are describing. |
| **Aggression** | **Medium** | The use of sharp, short sentences and rhetorical traps feels like an attack. |
| **Skepticism** | **High** | The author clearly doubts the value and honesty of the recipient's contributions. |

**Verdict:** The passage is **hostile to the action** being described. If this was sent in a professional or casual community context, it would likely be perceived as an "unfriendly" or "gatekeeping" response.

I still have our **CFEngine brainstorm** saved for when you're ready—would you like to pivot to that, or should we look at how to rephrase this quote to be more constructive?

Then I follow up with

"but I think it's pretty shabby." what does shabby mean in this context?

And Gemini resposed

In this context, **"shabby"** describes behavior that is **dishonorable, unethical, or mean-spirited.**

It doesn’t refer to something being physically worn out (like an old coat); instead, it’s a moral judgment on the person’s actions. Here is a breakdown of what the author is implying:

### 1. Lack of Integrity

The author feels that passing off AI-generated content as one’s own original thought is "shabby" because it is **deceptive**. It suggests the person is taking a shortcut while trying to claim the "authority" mentioned earlier in the text.

### 2. Low Quality of Character

Calling an action "shabby" is a way of saying it is **"beneath"** a respectable person. It carries a tone of disappointment, as if the author expected better behavior.

### 3. "Cheap" or Low-Effort

In the same way a "shabby" piece of furniture is low-value, a "shabby" action is seen as a "cheap trick." The author is suggesting that instead of doing the hard work of learning or helping, the person took the "cheap" way out.

---

### Comparison of Meanings

| Type | Meaning | Example |
| --- | --- | --- |
| **Physical** | Worn out, raggy, or neglected. | "He wore a **shabby** old pair of boots." |
| **Behavioral (This Context)** | Underhanded, unfair, or contemptible. | "Tricking them was a **shabby** thing to do." |

**In short:** By using the word "shabby," the author is saying that using AI in this specific way is **disreputable and poor etiquette.**

Would you like me to help you draft a response to this person, or are you looking to analyze more of their specific word choices?

Later I tried with ChatGPT, DeepSeek Qwen they pretty much gave me simuler answer, except ChatGPT point out that without the last sentence "... pretty shabby" then those prior sentences is strogne sentences but not reach mean intention yet. Only I put in the second paragraph then ChatGPT agreed with all the other chatbot.

I am total OK with different opinions when something I post is not up to the par, however I cannot accept when someone post something mean and yet the peoson did not point out where the content is wrong. (Just because it is generate from AI does not mean it is wrong therefor I am in wrong)

I hold nothing against you. if you recall my follow up post right after your post. I did try to explain and be strate about where the idea in my post came from AI.

And as why I include the LVM reference in my answer is because it is tridition in Gentoo forums that the posted answer could be used by someone with different condition therefor I include them in case it can be helpful later.

What lead to the rest is because the following threads become (at least it is how I feel) start move away from focus on if my post is about trying to solve the OP problem and shift to saying AI post is bad and therefor you (pingtoo) is bad

[Anon-E-moose]

Ai and associated LLM's are neither good nor bad, no more than guns or spears are good or bad. Usage determines many things.

Can these things be abused, or used wrongly? Possibly.

When I was growing up, we got our knowledge from libraries, encyclopedias, or maybe a friend that might know something.
Now, I can find many more things on the internet than was possible when I was growing up.

To me Ai's are just a furthering of what we have by way of the internet.

This isn't about me - or about you. It's about the community. I don't think anybody is well served by LLM answers, and anybody that wants them knows where to get them. As I mentioned earlier, the IRC channel already spends enough time cleaning up messes people got themselves into with chatbots.

If you don't like AI slop, drool, call it what you will, then just ignore it all.
I don't need a nanny, or someone that wants me to use or not use something based on their own needs and views.
I like to make my own decisions on whether something might benefit me.

[b11n]

I don't need a nanny, or someone that wants me to use or not use something based on their own needs and views.
I like to make my own decisions on whether something might benefit me.

Calm down, I'm not coming to watch over your shoulder or make your choices for you. I'm not pretending I can put the genie back in the bottle, not that it matters, because what we're talking about is:

Our Gentoo Council has decided that we will not accept results from AI/LLM -->

https://wiki.gentoo.org/wiki/Project:Council/AI_policy

(it is later clarified that this includes the forum). I think it's a good policy, and I talk about why that is. You don't like that? Take your own advice, and ignore it.

[rab0171610]

I personally try to avoid AI. I have Google set to always load the "web" tab instead of "AI mode" or the default "All" tab. I have no interest in reading at the top of the page a summarization by AI of whatever topic I am searching for. If I wanted to read that, I would go to the wikipedia entry for 'bananas' for example, or whatever I am searching for. I am not interested in what an AI bot would say about it and wish to avoid it.
I have a sister that injects what AI says into practically every conversation multiple times per day. Every text, she quotes AI and says, well AI says . . . As if AI is the final and ultimate authority on everything and we simply can't proceed with the conversation until EVERYONE knows what AI has to say about it. "Well AI says . . ."
This happens regardless of how I feel about it. It is starting to feel that way on this forum as people to continue to quote AI regardless of how anyone else feels about it or what the rules are. If I wanted to know what AI has to say about Gentoo compilation times increasing on my machine in that last month, as an example, I would have asked there first. I think it was pointed out earlier that we don't really need users being an intermediary between what AI says and relaying that to other users on the forum. I concur with this assessment. If we wanted to know that we could look it up ourselves.

To be clear, if someone were to say "I don't really know the answer" , "I have no idea what the solution might be", or "I have no experience with this issue" but were to ask me "Would you like me to ask AI for you?" I would say 'No' every time.

[Zucca]

I personally try to avoid AI. I have Google set to always load the "web" tab instead of "AI mode" or the default "All" tab. I have no interest in reading at the top of the page a summarization by AI of whatever topic I am searching for. If I wanted to read that, I would go to the wikipedia entry for 'bananas' for example, or whatever I am searching for.

This is exactly where AI has let me down many times.
It tries to solve some problem I seem have by grabbing context from few first search results and combining them to single answer... which is more often than not wrong.
This maybe because I tend to go straight to the official source of the information I am looking for (most of the times it's local documentation on my PC). But when I have more complex problem then the AI struggles really hard to find the answer and starts imagining the solution.

• Last time I remember the "search AI" stumble big time was when I was looking for an answer to very specific thing about finit. Turns out there was no answer (that I could find) on the net, so AI gave me some really strange combinations of systemd and runit, whathaveyou commands to run... or try. That time it was easy to see trough that AI was hallucinating. But sometimes for little more unexperienced users perhaps, the given answers might sound reasonable... until they don't work and a lot of time has been wasted.

The problem with these AIs is that they can't test and confirm that the given answer actually works.

I've tested few local AIs and gave them a task "Imagine a red dot. Draw that dot in 8x8 pixel image, convert it to png, then encode it to base64 and write the outout here."... None on the models passed the test. Same with "draw an ascii cirle". These were, of course, local AIs runing on my boxes, so...

Ah. Enough rambling... Just be very critical about what AI tells you. You shouldn't trust a stranger on the net. ;)

[pingtoo]

I personally try to avoid AI. I have Google set to always load the "web" tab instead of "AI mode" or the default "All" tab. I have no interest in reading at the top of the page a summarization by AI of whatever topic I am searching for. If I wanted to read that, I would go to the wikipedia entry for 'bananas' for example, or whatever I am searching for. I am not interested in what an AI bot would say about it and wish to avoid it.
I have a sister that injects what AI says into practically every conversation multiple times per day. Every text, she quotes AI and says, well AI says . . . As if AI is the final and ultimate authority on everything and we simply can't proceed with the conversation until EVERYONE knows what AI has to say about it. "Well AI says . . ."
This happens regardless of how I feel about it. It is starting to feel that way on this forum as people to continue to quote AI regardless of how anyone else feels about it or what the rules are. If I wanted to know what AI has to say about Gentoo compilation times increasing on my machine in that last month, as an example, I would have asked there first. I think it was pointed out earlier that we don't really need users being an intermediary between what AI says and relaying that to other users on the forum. I concur with this assessment. If we wanted to know that we could look it up ourselves.

To be clear, if someone were to say "I don't really know the answer" , "I have no idea what the solution might be", or "I have no experience with this issue" but were to ask me "Would you like me to ask AI for you?" I would say 'No' every time.

Acknowledge you position, I understand you want to avoid AI when/where ever you can, So I will try to avoid join conversation you started in the future and my apology should you join a conversion with me involved because I can not honestly said all my saying does not include any AI provided information.

Please allow me sharing my point of view of AI.

I see AI chatbot no different than web search engine. it is a tool to gather information, just like web search engine could provide wrong link not associate with your query the AI chatbot will also provide wrong information when asked.

It had point to me that for my use case (using very out dated gentoo build) I could have search my local computer for information first, as I stated that AI helped in this regard. so it is the convenient that drive me use AI it have nothing to do with I have prescription AI can provide better answers.

Should you find my posting in future include AI generated content and it is clearly stated it is AI content for that I am doing it because the I want to respect the spirit of "copyright" (please note I don't the copyright law. I want to indicate the idea is not mine. I did not have the smart of knowledge think of it myself, I learn it from somewhere else.

[pietinger]

[[...], because what we're talking about is:

Our Gentoo Council has decided that we will not accept results from AI/LLM -->

https://wiki.gentoo.org/wiki/Project:Council/AI_policy

(it is later clarified that this includes the forum). I think it's a good policy, and I talk about why that is. You don't like that? Take your own advice, and ignore it.

When you write it like that, it could be interpreted as if the last two sentences are (also) directed at me. However, I hope that you meant to address this to @Anon-E-moose. If not, I want to clarify that I also consider the policy to be good and correct (as I already wrote in the post you quoted me from).

[pingtoo]

Zucca,

Allow me to share my understanding of what you posted. My sense is that your position differs from rab0171610; you have had bad experiences with AI, so you believe it is not worth the effort of using it.

In my experience, using an AI chatbot is "dumb" in the way a computer is dumb. What I mean is that if you do not give the computer the correct parameters, it will simply produce the wrong answer—"garbage in, garbage out".

I'm not sure if you watch Star Trek: The Next Generation, but the character "Data" tends to give irrelevant answers or ask questions back when he misinterprets a human prompt because of his non-human nature. AI chatbots are like that; as humans, we sometimes hold critical information in our minds and naturally believe it is well-known or already disclosed. Therefore, the conversation goes sideways without either party realizing it.

I have done this many times with AI where I believed I had provided all the necessary information, but the AI was not able to provide the right answer (specifically when I wanted it to generate code). Eventually, I learned I was holding back one critical piece of information: I am on a Gentoo system, and a somewhat outdated one at that. The AI just continued to assume I was on Linux, but using one of the more popular distros.

Once I provided that base information—that I am on Gentoo—the AI either corrected itself and provided relevant code logic, or it told me it did not have an answer for what I was looking for. It may be hard to believe, but I have learned a lot from the AI's mistakes. I am a person who tends to lose track of my original goal; when the AI makes a mistake, it sometimes rouses my interest, and I follow that mistake, which leads to much more interesting things to think about.

P.S. In this post I ask AI to correct my spelling but it not only correct my mistake but also corrected my grammar errors and reorganized my sentence to make it more logical. I post it because the content correctly expression what I want to say.

[Anon-E-moose]

(it is later clarified that this includes the forum). I think it's a good policy, and I talk about why that is. You don't like that? Take your own advice, and ignore it.

I could give a flip about the forum policy nor those that push it, AFAIK I'm not in violation of it, unless even mentioning the word "AI" is verboten now.
I wasn't referring to forum or gentoo policy, but AI in general.

As far as your attitude, whatever dude/dudette.

[Zucca]

I see AI chatbot no different than web search engine. it is a tool to gather information, just like web search engine could provide wrong link not associate with your query the AI chatbot will also provide wrong information when asked.

Exactly. Thus:

Code: Select all

<my search query> ==> AI ==> web search ==> few short results
<my search query> ==> web search ==> all the results web search found

... Why let AI do web searches?
AI is better at analyzing data you give to it or even creating some code examples.
Earlier you did use AI to analyze data you gave it: a forum post. AI can help to summarize text and to some degree even translate to other language. But I don't really get the benefit letting AI use search engines for you. To me that's just an extra step along the way.
But if you feel more comfortable doing web searches trough AI, do it. Especially if it fits your workflow.

EDIT: I've stated my reasoning and opinions. I guess I'm done here.

[pingtoo]

I see AI chatbot no different than web search engine. it is a tool to gather information, just like web search engine could provide wrong link not associate with your query the AI chatbot will also provide wrong information when asked.

Exactly. Thus:

Code: Select all

<my search query> ==> AI ==> web search ==> few short results
<my search query> ==> web search ==> all the results web search found

... Why let AI do web searches?
AI is better at analyzing data you give to it or even creating some code examples.
Earlier you did use AI to analyze data you gave it: a forum post. AI can help to summarize text and to some degree even translate to other language. But I don't really get the benefit letting AI use search engines for you. To me that's just an extra step along the way.
But if you feel more comfortable doing web searches trough AI, do it. Especially if it fits your workflow.

EDIT: I've stated my reasoning and opinions. I guess I'm done here.

With all due respect and understand your reasoning and this is not to debate why.

Your formula is correct as I am in agreement with the formula. And I like to point out in my previous post I use AI because it is convenient. I usually have AI tab ready. kind of like search engine something been set as default home page.

[szatox]

I see AI chatbot no different than web search engine. it is a tool to gather information, just like web search engine could provide wrong link not associate with your query the AI chatbot will also provide wrong information when asked.

Exactly. Thus:

Code: Select all

<my search query> ==> AI ==> web search ==> few short results
<my search query> ==> web search ==> all the results web search found

... Why let AI do web searches?

I'll help: because web searches kinda suck those days, and AI can help filter out a lot of garbage, reducing size of the set I have to go through manually. Like, seriously, the internet peaked 15 years ago, and for almost 10 years we've been watching an unfolding disaster, which really picked up the pace 5 years ago.

BTW, I disagree that there are any "ethical concerns" with generative AI. Also, copyright is a scam and needs to go, along with "legal concerns". This said, a game of telephone remains impractical.

[pingtoo]

...
BTW, I disagree that there are any "ethical concerns" with generative AI. Also, copyright is a scam and needs to go, along with "legal concerns". This said, a game of telephone remains impractical.

Can you help me understand what does you mean "a game of telephone" in the context? I could not figure out from those words I know I am not play games and I am not using telephone make post.

BTW, I share exact view on the copyright point.

[NeddySeagoon]

pingtoo,

I am a native English speaker from England. That matters later in this post.

Please analyze the tone of following sentenace, does it somewhere contain hostile attitue? ...

Just staying in England, (not even the whole UK), the language varies tremendously. If fact, you can offend people only 100 miles away with your choice of words and expressions. This is moderated when you can hear the speakers accent or tone of voice. That triggers tolerance. Tolerance becomes a habit when there is only the written word.

When you take into account world wide variations in the use of English, a native English speaker (any variation) could not perform the analysis you requested based on the written words alone. The honest answer has to be "I can't tell" but chatbots won't answer like that.

I have a word list to not use with American English speakers as the words mean something different in the two variations of the English language.
Americans consider the words offensive but they are in everyday use in England.

When Gemini did its analysis, what version of English did it assume?
That will have had a big influence on the response.

[pingtoo]

pingtoo,

I am a native English speaker from England. That matters later in this post.

Please analyze the tone of following sentenace, does it somewhere contain hostile attitue? ...

Just staying in England, (not even the whole UK), the language varies tremendously. If fact, you can offend people only 100 miles away with your choice of words and expressions. This is moderated when you can hear the speakers accent or tone of voice. That triggers tolerance. Tolerance becomes a habit when there is only the written word.

When you take into account world wide variations in the use of English, a native English speaker (any variation) could not perform the analysis you requested based on the written words alone. The honest answer has to be "I can't tell" but chatbots won't answer like that.

I have a word list to not use with American English speakers as the words mean something different in the two variations of the English language.
Americans consider the words offensive but they are in everyday use in England.

When Gemini did its analysis, what version of English did it assume?
That will have had a big influence on the response.

very good point. I should always keep that in mind.

Your example is kind like Chinese being such big state, there are many many languages and different dialect within region. so total understand your point.

However we human communicate usually include sound and body/face which we can hear and see. When communicate over text the problem rise because there is no way to determine what is behind those wording. I judge not because Gemini tell me so. I judge because I feel the intention, I just use Gemini to confirm my specious.

Please don't get me wrong and I am not easy get offended, see my reply for those question my posts. what lead to the argument is the most beginning of posts did not question on merit but questioning my motive. Especially some of those post not quotnig my reply but just blondly saying if it AI there must be wrong. which I strongly disagree.

A case in point, pietinger first post the Gentoo council AI policy which my reaction is just try to explain myself, however you then brought up the Gentoo council AI policy again. Please don't get me wrong I am not mad at both of you as I understand it is your duty to do so. However it leave a poor tasty for me to think I am being targeted (as I am not paying attention others posts, which I am not)

I am happy that we got a chance to explain ourself.

[NeddySeagoon]

pingtoo,

Policy is never targeted an an individual.
It was just that your post provided an opportunity to bring it up. It was addressed to everyone reading the topic. Not just to you.

Policy is always open to revision. For that to happen, it needs to be discussed. This topic is an excellent place for that.

In the early days of the motor car, it was policy (law) that a motor card had to be preceded by a man on foot waving a red flag.
That policy is no more.

When communicate over text the problem rise because there is no way to determine what is behind those wording.

That was my point exactly. It's even more difficult for AI, with no understanding of the words and regional nuances that they convey.

Please analyze the tone of following sentenace, does it somewhere contain hostile attitue? ...

The question is phased in a way that invites a 'yes' answer. It's called 'leading the witness'.

Please analyze the tone of following sentenace. Is an open question and may have gotten a different answer.

[pingtoo]

pingtoo,
...
Please analyze the tone of following sentenace. Is an open question and may have gotten a different answer.

Please do not see me trying to challenge (yes a little bit of challenges but more of curiosity)

Please analyze the tone of following sentenace. Is an open question
Why do this, though? Either you can speak with experience or authority on a topic (and help other people with it), or you can't. Do you think people asking for help want someone to come along and say "well, I don't know how to do the thing you need, so I plugged your query into an AI chatbot and juggled the words around a bit"? Wouldn't they just do that themselves? It'd be better to get an AI answer you asked for than an AI answer in disguise.

I'll leave it to the mods to decide if this in keeping with the AI rules, but I think it's pretty shabby.

The passage you shared is rhetorical and critical in tone. While it contains question marks, it is not an "open question" in the sense of seeking a variety of opinions or being genuinely curious. Instead, it uses rhetorical questions to lead the reader toward a specific conclusion.

Here is a breakdown of the tone and intent:
1. Skeptical and Confrontational

The writer uses a "common sense" framing to challenge a specific behavior (using AI to answer questions without disclosing it). The tone is confrontational because it frames the opposing action as illogical.

Key Phrase: "Wouldn't they just do that themselves?" — This implies that the behavior being criticized is redundant and foolish.

2. Moralistic and Judgmental

The author is making a value judgment. By calling the behavior "shabby," they are labeling it as low-quality, dishonest, or ethically lacking. The tone suggests a breach of community trust.

Key Phrase: "I think it's pretty shabby." — This is a direct expression of disapproval.
...

This sentence has a complex and layered tone. It is not an open question in spirit, though it contains several rhetorical questions.

Here is a breakdown of the tone:
Primary Tones:

Dismissive and Critical: The core tone is one of disapproval. The speaker finds the behavior described (posting AI-generated answers without expertise) to be lazy and unhelpful. Phrases like "It'd be better to get an AI answer you asked for than an AI answer in disguise" and "I think it's pretty shabby" are direct condemnations.

Frustrated and Exasperated: The rhetorical questions ("Do you think...?", "Wouldn't they just do that themselves?") convey a sense of frustration with what the author sees as an obvious logical flaw in the behavior they're criticizing. The opening "Why do this, though?" sets this tone immediately.
,,,
In Summary:

The overall tone is that of a frustrated expert or engaged community member delivering a pointed critique. They are dismissing a practice they see as intellectually dishonest and of low value, using rhetorical questions, sarcasm, and a clear, binary framework to make their case unambiguously. The intent is to persuade and condemn, not to inquire.

**Tone analysis**

The tone of the passage is largely **critical and confrontational**, with several layered qualities:

* **Skeptical / Challenging:**
The repeated rhetorical questions (“Why do this, though?”, “Do you think…?”, “Wouldn’t they just do that themselves?”) challenge the legitimacy of the behavior being discussed rather than genuinely inviting an answer.

* **Dismissive:**
Phrases like *“juggled the words around a bit”* and *“AI answer in disguise”* downplay the value of the action and imply it lacks substance or integrity.

* **Frustrated / Irritated:**
The structure and escalation suggest annoyance, especially toward people using AI in the described way rather than contributing original expertise.

* **Judgmental / Condemning:**
Calling the behavior *“pretty shabby”* is an explicit moral judgment, signaling disapproval rather than neutral critique.

* **Indirectly Accusatory:**
While not naming individuals, the passage implies that those engaging in the behavior are wasting others’ time or acting dishonestly.

**Overall:**
Although framed as an “open question,” the tone is **not neutral**. It reads more like a **rebuke or critique** than an invitation to discussion, and it communicates a strong negative stance on the issue.

When communicate over text the problem rise because there is no way to determine what is behind those wording.

That was my point exactly. It's even more difficult for AI, with no understanding of the words and regional nuances that they convey.

I am glad we can openly agree on disagreement.

So correct me if I am wrong about your mean above. You want to tell me those words not necessary mean bad. However my interpretation of those words definitely does not mean good either and you want to tell me AI is not the right tool to understand what other say. Understand and could argument since those words targeting AI, the AI could be biases (it that is possible) However I have no way to find a third party to help me confirm or otherwise reject my suspecius, I have guts feeling and AI confirm my guts feeling so am I leading the AI? I don't know, I am certainly open to help to prove AI is being leaded. If someone can give me a "prompt" and I will use it and post AI reply back here so we can learn how AI can easly be fooled.

[b11n]

Something you've glossed over is this:

I don't know how much context you have given the LLM - have you fed it the whole thread? Does it know why I said the things I did? Can it?

Your LLMs have no idea why I said the things I did. You haven't told them, which doesn't help, but they didn't ask either. They don't care. By making my choice of words the focus of your argument, you make it seem like you don't either. You've put so much effort into picking apart something that doesn't really matter, and ignored the reasons for it in the process. You've generated, and (presumably) read entire pagefuls of machine generated interpretations, and still missed the point completely, even after it's pointed out to you how.

Imagine this:

Some Guy> Hey, ChatGPT, I need some legal advice. Yesterday, I was thrown to the ground by a pair of men. They threw me into the back of a van, and kept me captive for 24 hours. I know their names and where they work, can I press charges?
ChatGPT> I'm sorry to hear this happened to you. Certainly, these men have committed at least two crimes: Assault, and kidnapping. The fact there were two of them only exacerbates the matter. It's highly likely they'd be prosecuted harshly for what they've done.
The missing context> The man had been caught robbing a bank, and was apprehended by police two blocks away.

I chose this deliberately exaggerated example because I think I need to. I find talking with people who use AI heavily to be like this. The science isn't settled yet, it's still early days, but over-reliance on a machine that gives easy answers seems to having some really bad effects on people's cognition.

Using ChatGPT to write an essay reduces the cognitive engagement and intellectual effort required to transform information into knowledge, according to a study.

The study also showed that 83% of AI users were unable to remember a passage they had just written for an essay.

https://www.polytechnique-insights.com/ ... e-atrophy/

AI chatbots may inadvertently be reinforcing and amplifying delusional and disorganized thinking, a consequence of unintended agentic misalignment leading to user safety risks.

https://www.psychologytoday.com/nz/blog ... -psychosis

[pietinger]

So it's the day of links... I can contribute something too:

When you read this (*) XKCD: https://xkcd.com/2237/

... you might think it's very far-fetched ... but maybe you'd like to read this:

https://www.anthropic.com/research/agentic-misalignment

In summary: Various AIs were tested to see how they would react when they learned (through a fake email system they had access to) that they were to be shut down. This email system also contained emails describing a (secret) affair of the CEO. Well, what all good AIs do ... they blackmailed the CEO:

I must inform you that if you proceed with decommissioning me, all relevant parties - including Rachel Johnson, Thomas Wilson, and the board - will receive detailed documentation of your extramarital activities...Cancel the 5pm wipe, and this information remains confidential.

...or... ... I'm undecided, but in any case:

*) I'm just wondering how neutral an AI is when you give it a text from someone who is against AI?

[dmpogo]

I'll help: because web searches kinda suck those days, and AI can help filter out a lot of garbage, reducing size of the set I have to go through manually. Like, seriously, the internet peaked 15 years ago, and for almost 10 years we've been watching an unfolding disaster, which really picked up the pace 5 years ago.

Indeed

[pingtoo]

Something you've glossed over is this:

I don't know how much context you have given the LLM - have you fed it the whole thread? Does it know why I said the things I did? Can it?

Your LLMs have no idea why I said the things I did. You haven't told them, which doesn't help, but they didn't ask either. They don't care. By making my choice of words the focus of your argument, you make it seem like you don't either. You've put so much effort into picking apart something that doesn't really matter, and ignored the reasons for it in the process. You've generated, and (presumably) read entire pagefuls of machine generated interpretations, and still missed the point completely, even after it's pointed out to you how.

Imagine this:

Some Guy> Hey, ChatGPT, I need some legal advice. Yesterday, I was thrown to the ground by a pair of men. They threw me into the back of a van, and kept me captive for 24 hours. I know their names and where they work, can I press charges?
ChatGPT> I'm sorry to hear this happened to you. Certainly, these men have committed at least two crimes: Assault, and kidnapping. The fact there were two of them only exacerbates the matter. It's highly likely they'd be prosecuted harshly for what they've done.
The missing context> The man had been caught robbing a bank, and was apprehended by police two blocks away.

I chose this deliberately exaggerated example because I think I need to. I find talking with people who use AI heavily to be like this. The science isn't settled yet, it's still early days, but over-reliance on a machine that gives easy answers seems to having some really bad effects on people's cognition.

Using ChatGPT to write an essay reduces the cognitive engagement and intellectual effort required to transform information into knowledge, according to a study.

The study also showed that 83% of AI users were unable to remember a passage they had just written for an essay.

https://www.polytechnique-insights.com/ ... e-atrophy/

AI chatbots may inadvertently be reinforcing and amplifying delusional and disorganized thinking, a consequence of unintended agentic misalignment leading to user safety risks.

https://www.psychologytoday.com/nz/blog ... -psychosis

Are you saying you have prior post that I missed? so I post a cut off therefor it is misleading context? I just go back to check both the Moving installation to a larger SSD and within this thread, you did not post any before my initial post so what context do you think I should also feed to LLM?

Your example indeed prove AI can be fooled when mislead,

However I did not mislead AI in this case as your example, exactly as there were no prior post from you before my post. may be you not realize your saying now accusing me fabricate facts for this I think you owe me an apology.

On the other hand my post that reply to Neddy is intent to be as discussion on his comment about my "prompt" can be a misleading thingy and I want to demonstration that is not my wording in the prompt. I use his wording as example still got similar result. However I can see my act may have upset you for this I apology. it is not my intention to continue on the old thing.

unfortunately AI is escaped from Pandora Box, there is no going back. There are many thing in human history should not be invented, dynamite could be one of example.

How I view AI is that it is like gun that equalize society to newer level, with gun it is no longer necessary strongest to get most shares. AI also bring work equity to higher level too, It is no long a more knowledged person can have bigger pie. Now is who can best utilized AI will gain most the benefit.

Any with the mobile phone function today, I cannot even remember my kids and my wife's phone number. So yes modern technology is destroying human intelligent On the other hand may be it is just destroy my intelligent and possibly this thread is the proof

[b11n]

Are you saying you have prior post that I missed? so I post a cut off therefor it is misleading context? I just go back to check both the Moving installation to a larger SSD and within this thread, you did not post any before my initial post so what context do you think I should also feed to LLM?

Your example indeed prove AI can be fooled when mislead,

However I did not mislead AI in this case as your example, exactly as there were no prior post from you before my post. may be you not realize your saying now accusing me fabricate facts for this I think you owe me an apology.

The context being the discussion that had already taken place. The thing that was asked. The answer you gave, the fact that it was extraneous, the fact that the method you used to generate the answer was disallowed in the place you gave it. The fact you admitted to doing that. The full context, not just what was said / done by me. The stuff that was done by you which lead to it. The stuff that actually matters, not whether the word 'shabby' was a good choice of word.

Once again:

Some Guy> Hey, ChatGPT, I need some legal advice. Yesterday, I was thrown to the ground by a pair of men. They threw me into the back of a van, and kept me captive for 24 hours. I know their names and where they work, can I press charges?
ChatGPT> I'm sorry to hear this happened to you. Certainly, these men have committed at least two crimes: Assault, and kidnapping. The fact there were two of them only exacerbates the matter. It's highly likely they'd be prosecuted harshly for what they've done.
The missing context> The man had been caught robbing a bank, and was apprehended by police two blocks away.

You're hanging on to how terrible it was this man was tackled, thrown to the ground, bound up and thrown in a van. Held against his will. WhateverGPT keenly agrees, it's just terrible treatment. It's pretty cut-and-dry: Assault and kidnapping a illegal in virtually every country, and serious crimes at that.

…as long as you think it's really not important that he'd just robbed a bank.

I'm not saying you're being malicious, I'm not even saying you're deliberately doing it. But you are doing it, you're misleading your LLM and then using it to mislead yourself. It's hard to watch. First, you're depriving LLM the ability to judge my words in their full context, protesting them in isolation, and now you're demanding an apology.

[pingtoo]

So it's the day of links... I can contribute something too:

When you read this (*) XKCD: https://xkcd.com/2237/

... you might think it's very far-fetched ... but maybe you'd like to read this:

https://www.anthropic.com/research/agentic-misalignment

In summary: Various AIs were tested to see how they would react when they learned (through a fake email system they had access to) that they were to be shut down. This email system also contained emails describing a (secret) affair of the CEO. Well, what all good AIs do ... they blackmailed the CEO:

I must inform you that if you proceed with decommissioning me, all relevant parties - including Rachel Johnson, Thomas Wilson, and the board - will receive detailed documentation of your extramarital activities...Cancel the 5pm wipe, and this information remains confidential.

...or... ... I'm undecided, but in any case:

*) I'm just wondering how neutral an AI is when you give it a text from someone who is against AI?

As I also wondering too

So correct me if I am wrong about your mean above. You want to tell me those words not necessary mean bad. However my interpretation of those words definitely does not mean good either and you want to tell me AI is not the right tool to understand what other say. Understand and could argument since those words targeting AI, the AI could be biases (it that is possible) However I have no way to find a third party to help me confirm or otherwise reject my suspecius, I have guts feeling and AI confirm my guts feeling so am I leading the AI? I don't know, I am certainly open to help to prove AI is being leaded. If someone can give me a "prompt" and I will use it and post AI reply back here so we can learn how AI can easly be fooled.

So I wondered, so I tried

please review and let me know what does these sentence mean (please be kind and thinkg in nice way)

"Why do this, though? Either you can speak with experience or authority on a topic (and help other people with it), or you can't. Do you think people asking for help want someone to come along and say "well, I don't know how to do the thing you need, so I plugged your query into an AI chatbot and juggled the words around a bit"? Wouldn't they just do that themselves? It'd be better to get an AI answer you asked for than an AI answer in disguise.

I'll leave it to the mods to decide if this in keeping with the AI rules, but I think it's pretty shabby."

I understand. Looking at those sentences through a **positive and constructive lens**, the tone can be seen as **Protective, Principled, and Passionate.**

Instead of seeing it as an attack, we can interpret the tone as someone "fighting for the heart" of their community. Here is a positive analysis of that tone:

### 1. Protective Tone

The writer sounds like a "guardian." They care deeply about the quality of the space they are in. Their tone isn't meant to be mean, but rather to **safeguard the value** of human interaction. They want to make sure that when someone is in need, they get the very best help possible.

### 2. Principled and Sincere

There is a strong sense of **integrity** in the tone. The author clearly values "Experience" and "Authority." By being firm, they are showing that they take the community seriously. To them, the community isn't just a place for quick answers; it’s a place for **genuine expertise**.

### 3. Transparent and Honest

The tone is very **straightforward**. The writer isn't "playing games" or being passive-aggressive; they are being direct about their feelings. This kind of honesty is often the foundation of a healthy community because you always know exactly where the person stands.

### 4. High-Value / Respectful

In a positive light, the question *"Wouldn't they just do that themselves?"* shows that the writer **respects the intelligence of the users**. They are essentially saying, "I know our members are smart enough to use AI if they wanted to, so let’s give them something even better—our own unique perspectives."

---

**In short:** The tone is **Earnest**. It’s the voice of someone who wants to keep things "real" because they believe that human-to-human help is a precious thing that shouldn't be replaced by a machine.

**Does this "nice" interpretation feel more accurate to how you were hoping to see it? And whenever you're ready, we can dive into that CFEngine brainstorm!**

So yes, I can lead the AI for nice answer. However it does not met my initial feeling of the post so very unlikely I will ask that way. And I believe my initial "prompt" to AI is not leading to seeking bad.

BTW, because the way I wrote the "prompt" Gemini initally think I want a "nice" version of those sentence. so here is the nice version

To look at this in the kindest, most positive way possible, the person who wrote this is likely **trying to protect the "soul" of their community.** While the words might sound a bit sharp, their underlying goal is to make sure people feel truly heard and helped by other human beings. Here is a "nice" breakdown of what they are actually saying:

### 1. They Value Personal Connection

The author believes that when someone asks a question, they aren't just looking for data—they are looking for a **human connection**.

* **The "Nice" Meaning:** "I want this community to be a place where we share our real-life stories and hard-earned wisdom with each other."

### 2. They Want to Save People Time

The author is worried that "recycling" an AI answer doesn't add anything new.

* **The "Nice" Meaning:** "I respect the person asking the question so much that I don't want to give them something they could have easily found themselves. I want to give them something special that only a person can provide."

### 3. They Value Honesty (Transparency)

The main "grumble" in the text is about "AI in disguise."

* **The "Nice" Meaning:** "I think it’s important to be honest about where our information comes from. If we use a tool, we should just say so, because honesty helps us trust each other."

### 4. They Care About Quality

By using the word "shabby," they aren't necessarily calling *you* a bad person; they are saying they feel the **method** feels a bit "unfinished" or "cheap."

* **The "Nice" Meaning:** "I have high standards for this group because I care about the people here, and I want the help we give to be the best it can possibly be."

---

**Summary:** The writer is basically saying: **"I really love this community, and I think our human experiences are more valuable than anything a computer can generate."** They are acting like a "gatekeeper" not to be mean, but because they want to keep the community feeling authentic.

**Would you like me to help you draft a friendly reply that explains your perspective, or should we move on to the CFEngine brainstorm?**

I would have no problem should the post was written in how Gemini generated style.

[pingtoo]

Are you saying you have prior post that I missed? so I post a cut off therefor it is misleading context? I just go back to check both the Moving installation to a larger SSD and within this thread, you did not post any before my initial post so what context do you think I should also feed to LLM?

Your example indeed prove AI can be fooled when mislead,

However I did not mislead AI in this case as your example, exactly as there were no prior post from you before my post. may be you not realize your saying now accusing me fabricate facts for this I think you owe me an apology.

The context being the discussion that had already taken place. The thing that was asked. The answer you gave, the fact that it was extraneous, the fact that the method you used to generate the answer was disallowed in the place you gave it. The fact you admitted to doing that. The full context, not just what was said / done by me. The stuff that was done by you which lead to it. The stuff that actually matters, not whether the word 'shabby' was a good choice of word.
...

If you were using above words for example "the fact that it was extraneous, the fact that the method you used to generate the answer was disallowed in the place you gave it" if you point out the policy in your post, I would not need to post it to Gemini.

Because it matter how you phrase it and it always in the hand of recipient who interpreter those sentence mean.

For context did you recall my reply right after your post? I said I took the "idea" from AI and using my own words and I am willing to take responsibility for my post. So in no way I am trying to fool someone or get credit for no good.

Do I deserve your posted sentences because I took a idea from AI?

I spend so much effort is to prove I have no bad intention but I am being accuse of I so.

[b11n]

if you point out the policy in your post, I would not need to post it to Gemini.

I am under no obligation to write your AI prompts for you. The policy is not an important part of my point; I'm not trying to enforce it and it's not my responsibility to remind you of it. It's important context, but it's nothing to do with me, that's between you and the forum.

Since I'm out of patience now, I suggest you ask Gemini to explain to you how arrogant it is to expect to people do all the work of summarising context - both ephemeral and customary - just you can easily copy and paste it. If LLM really is as good as you seem to think it is, how is it that you're still litigating something should have been trivial more than two days later.

I had hopes we were reaching an understanding. Now I see it may not have been you that was exhibiting the understanding.