Linux

I've tried nothing and I'm all out of ideas.

I hate it when AI hallucinates an answer but I REALLY hate it when I try to ask a question and some smug know-it-all tells me I just want to be spoon fed the answers and didn't do any research on my own. Usually if I am to the point of writing a post asking for help I have tried the usual ways of finding answers (at least the ones that work for me - man pages decidedly do not) and I'm at the point where yes, I WOULD like to be just spoon-fed the answer if you know it (or at least a relevant link), thank you very much. More than once I have got REALLY pissed off at a guy like that and then usually I'm the one that gets censored or booted out of the forum, while the know-it-all goes on to find another victim to bully. Oh, and then there are the moderators who censor "low quality" posts (except from their friends and users they like).

So when AI came along, I found it much easier to ask it and try to work through the hallucinations than to deal with that, and I suspect a whole lot of other people felt the same, and that is one reason why sites like Stack Exchange saw participation fall off a cliff. And AI keeps getting better and (so far at least) has never once criticized me for not trying hard enough to find something on my own. I get some people want challenges in their lives, but not everyone does and even those who like to be challenged now and then don't always want it to be in the same areas of their lives. There are people who enjoy running in marathons but who would hate any friction at all when trying to use their computer.

AI scrapers only know this outdated information

While I have experienced this (quite a lot), it's much easier to spend five minutes figuring this out with an AI than it is to spend an hour trying to work that out by searching forums for answers.

I find AI way more up to date then searching.

Oh I hate how everything is a video. Sometimes I just want text so I can ship to relevant section.

Meh.... There have been a whole lot of braindead "pls do it for me" posts going around since the influx of windows refugees.

I'm not one to be a prick about answering questions to help out, and we all start somewhere, but "help me get steam and my LLM working on Kali linux" where ppl get in way over their heads is becoming a real problem. Worse, they will call the linux community toxic for suggesting they need more fundamentals first.

I interpret the post title as a general "try your best, and come to us when you're stuck".

Counterpoint: Aside from disabilities which specifically impact reading, why wouldn't someone want to read when it comes to their hobbies? A hobby is something one intentionally devotes time to, typically unpaid and nonprofessionally, because they enjoy it and they want to learn more about it. A large amount of my enjoyment is derived from learning more about the things I enjoy, so not wanting to consume that information makes no sense to me.

I can understand, for example, gaming as a hobby and wanting quick answers if one is jumping ship from Windows to Linux. Linux isn't the hobby there; just a means to an end. I'd still argue the gamer should develop some level of proficiency with their underlying OS. Otherwise it's like having trail riding as a hobby without any knowledge or tools to patch a tube, tension a chain or tighten a bolt. One might end up in a situation where they can't just get an instant answer. Investing a little time in the mechanics could keep a short ride from turning into a long trudge out with a bike over the shoulder.

In the context of "Linux", broadly, as a hobby - what even is that hobby if it isn't making an honest effort to learn broadly about various tools, the kernel, scripting and programming languages, and so forth? Linux always struck me as a hobby for people who collect hobbies. Or people that have "learning" as a hobby. It's why, while I'll probably never work a day of my life in IT, I know how to do some basic SQL queries, hit an API and parse the JSON, do a little scripting in Bash and Python, utilize a load of CLI tooling much more efficiently than any Windows GUI I've ever used, and so on. I'll never know it all but part of the fun is trying anyway.

"It's not a professional's job to read the manuals they need to know for their job unless I specifically tell them to" is an interesting take. A really stupid one but interesting non-the-less...

Sometimes people ask others instead of googling things because it offers an opportunity to socialize. We all know google is an option. I get what you're saying, but it's sad to see this described as a burden.

I agree that "RTFM" can be insensitive, and even mean. However, the place it comes from is genuine. It's nobodies job to tell you exactly what page to look at. If you've dug through the docs and still can't find your answer, make it explicit that you've searched the manual, and perhaps be explicit about parts you don't quite understand.

The whole "RTFM" thing was born from people asking for help when they obviously hadn't made a proper try themselves first.

often on linux, you can simply type man <command> and get TFM

i would agree with you on GUI apps, where it can be confusing.

What's worse, is if this is GNU-ware, there is a good chance the answer IS NOT IN THE MAN PAGE. I think it was bash or maybe gawk. I don't remember exactly, but I had a question that simply wasn't answered in each man page. GNU docs are absolute trash, written without any consideration for the audience.

there's a standardized man format? News to me. I thought developers just threw everything in at random order.

I can't count the number of times I swear I read every man page and I can't figure out how to structure my arguments--especially when they are nested or conditional.

I especially wish more man pages had common examples. Sometimes an example can say more than a paragraph of explanation.

I can relate : I'm a noob. I often don't understand the manual, so I look for forum posts that are clearer/ easier to follow/more directly related to my problem and most of the time I find some.

The problem with using an LLM for information is that you can ask chatgpt:

"please provide me with 3 different interpretations of the main function of the Linux command 'ls'

and you'll get what you ask for. An llm is an inappropriate tool to lookup accurate information.

I dunno man, I want my friends and family using Linux. I don’t think this is the right mentality.

People still converse over simple facts. What does the plant that grows cinnamon look like? Who won the Super Bowl in x year? Simple facts that could easily just be looked up, but people like talking to each other.

Even reading off a Wikipedia page to someone else, you get an opportunity to cater the tone, pace, and omissions / additional clarity to whomever you’re talking to.

The drop in Stack Overflow questions shows that if people can’t get helpful answers from other humans, they will get them from AI.

this is an absolutely toxic take of the issue. I took OP's statement as less of a "I won't read the manual" and more of a "I struggle to be able to read manuals"

There are many times I had read the manual, and then had to look up the issue further anyway because I either missed the poorly written section, or misunderstood what it was saying.

If you want a prime example of that, go look at ffmpeg and try to figure out how to select a specific language for subtitles on a video without looking it up online. its via -map as an advanced option, which is described as a parameter to extract specific streams (which also means they would need to map the video and the audio streams since including a -map removes every auto stream). but map doesn't tell you subtitle tracks are index:s. it does tell you that you can look at stream specifiers for valid search options, which does include s as a type, and lets you know that you can use m for metadata tagging, but you would need to make the connection that the type is s, and the meta data search flag would be m:language:langcode, and you need to make the connection the entire string has to be concated so its index:s:m:language:langcode For someone who is learning ffmpeg and video transcoding, that is not a very good setup. The stream specifiers give a few examples of what the potentials are but, the location where it specifies the types are in a different area than the one where it specifies the metadata keys. At that point just asking online or searching is way easier.

Note: this is just an issue I have see people come across because ffmpeg is one of the more complicated programs (the man page is over 2300 lines)

is it in the manual? yes. is someone who doesn't know how to use ffmpeg and is trying to learn it going to find it? that's debatable.

If I was in that situation, my next step would be googling it, and if I couldn't find it via searching, I would be reaching out to communities. At that point "RTFM" is useless to me.