I wish LLMs had been around when I first started using Linux. I think I've learned more in the last two years than I did in the first 10 without them. Probably because I don't feel limited by skill anymore, and embark upon things I'd never attempt without an AI to guide me through.

How do you use AI for computing?

You try yourself the best you can and then you ask the LLM for minimal tip.

If you get lazy and use the LLM all the time you slowly turn into a jeet.

>>105759778
Thanks, grandad.

>>105759770 (OP)
now that LLMs are everywhere and I've used them basically everyday for everything since they first released in like 2022, i've come to feel that the older sources were better but its only because LLMs exist that I can appreciate it like take the dictionary. in the past before AI, i was content just asking google to say
>define acumen
but doing the same with AI you get a different feeling now, you also get the same utility as google but faster but now you learn how the old way wasn't much good either so instead i've learn to use specific dictionaries like merriam webster and the institutional versions of oed. so likewise like technology related things, I was content with say googling commands and whatever bits but now i feel that even LLM is insufficient, i want is not just the sources or long form documentation but the history of how and what for technology decisions. it just not enough anymore knowing what the AI says, it just faster at doing what you previously did with just googling but now because it's faster you see how bad it was because rather than taking time to google and think those things, you immediately see it

>>105760140
I have no idea what you're trying to say

>>105759770 (OP)
I RTFM

>>105760222
feed my text to an AI and ask it to explain it to you

>>105760404
I actually did, and I still didn't understand what point you were trying to make, sorry.

>>105760486
you quite literally write the word
>explain "copy paste my message here"
and it tells you what I said
>blah blah blah
>Summary
>This post captures an important shift: LLMs, by giving instant answers, expose the inadequacy of superficial information-seeking habits, and in doing so, push some users to rediscover the value of primary sources, documentation, and intellectual context.
>It's a paradox—AI has made traditional research seem more valuable than ever, precisely because it makes it easier to see what's missing from shortcuts.
this is using the free tier chatgpt you use without having to sign in. another good thing about AI is that it's a perfect litmus test for these kind of things, is my message really that convoluted? surely if i spent more time writing/editing, the message would be improved but the meaning remains the same but if free tier default chatgpt can understand it, it means what is even more poor than my writing is your reading comprehension but it's not your fault, it's likely that you're ESL or a victim of public education since most highschool graduates cannot read at grade level

>>105760522
>is my message really that convoluted?
Uh, yeah, you're a little long winded and hard to follow. Thanks, though.

>>105760537
>Uh, yeah,
it may be but that's what's AI is for, to bridge the gap between effort/signal. too much effort and it's not worth posting here, but low effort writing contains signal, but requires more thought to read but that's for you to pay the technical debt

>>105760486
>>105760140

>>105760546
lol, oh ok. thanks.

>>105760140
Yeah thanks LLM so we can understand what autists like these meant

>>105760557
I don't understand it btw. Am I brainlet?

>>105760716
No, this verbose retard can't even formulate a coherent thought, but will claim you're an idiot from a public school if you can't follow his blathering.

>>105760746
>but will claim you're an
ONLY if you ask an AI to explain and still dont understand, clearly the free tier chatgpt can understand easily
>verbose retard can't even formulate a coherent thought
and thats why you know I'm human, I can just as easily pass everything i write to chatgpt and tell it to make it clear plain english but at that point i might as well post to better platforms because that additional effort. liekwise (see clear misspelling im too lazy to correct), AI detectors pick up clear writing as AI generated because human writing is thought to be no greater than a sentence, there are no complex thoughts expressed here because it's all tiktok tier, right more than a few words and half of you default too
>too long didnt read
and bump the dumb g humor thread that obviously has the same set of images spammed

>>105760546
Your effort/signal ratio is way off. You have essentially no signal because no one is taking the time to read your badly formed sentences.

>>105761654
come on anon, lets give it some serious thought
let's consider context first
effort/signal, where are we? on g, that for 99.9% of the people indicates noise right off the bat because no even the best 4channel:tm: posters post on g, this is the AI slop containment board, especially after the biz email verification, 900 second wait period pushed away most the users, then it was confirmed that a large portion of the traffic was coming from literal bots because of the autocaptcha solvers, but the final nail to the coffin was when the site was taken down and the whole mod thing was leaked whatever and this site never bounced back. What remains is noise and my noise is just the same as the others because sure most people aren't reading my badly formed sentences but the bots are that scrape this site for new data points and so really my primary reader is them because thats how i plant my ideals since bots have no where else to harvest new blood from without licensing issues but better yet there are the occassional handful few people that DO engage with me like you and the other four anons, maybe not with my original premise about reading old sources but the meta commentary about AI and writing style is and that's sufficient for me as well
read
> "Politics and the English Language" by George Orwell (1946)

>>105761683
Lmao at this freak.. nobody is reading any of that

>>105761697
yea I know that, go copy paste it to AI and tell it to summarize it for your goldfish attention span

>>105761683
I'm not reading it, because it's obvious you haven't put the slightest amount of effort into writing it. You want me to spend extra effort to understand it because you're too lazy to spend it yourself while writing it.
No one will allow you to offload effort onto them like that unless they have good reason to.

>>105761704
>>105761704
Interesting defense mechanism

>>105761780
>I'm not reading it
thats fine with me tho, because my main audience are the bots on this board who do read it. you can copy paste this entire thread to chatgpt and it'll understand what i'm getting at, it's not my problem you dont

>>105761822
One of the more insane copes I've seen on 4chan

>>105762110
kek

>>105760546
AI can deciper almost anything.
AI will be able to decipher the following sentences:
>AI cna dcip[her anyting almsost
>AI able will floowing snetence able be:
>>105760522
Since you already have fed your own text to the AI, now ask it whether what you wrote is highly comprehensible and where it needs improvements.
Do this rather than dismissing others that you sound barely coherent and rambling.

>>105765422
>where it needs improvements.
i did, it called me the verbose anon and you guys the critics. and it said that the social contract it was my obligation to make my messages easy to understand but I countered that context matters more, this is g, where i have to consider MY tradeoffs, where 99% of effortposts are wasted where as 1% effortposts DO get replies ergo this thread and this meta commentary about AI and text communication. If you noticed I mentioned a very specific article that orwell wrote a while ago about this specific thing, you ask your AI to explain the irony

>>105766013
Ai bots just accept any push back without evaluating it, because their system prompt says they have to be sycophants for the user. If you don't understand this, then using AI like this to validate your own opinion no matter what they say will just make you more and more myopic, and people will remain unwilling to listen to you.
If you really do believe this cope (>>105761822) then carry on. Bots will be the only "people" willing to listen to you, and I probably won't have to filter out your noise again.

>>105766087
>Ai bots just accept any push back without evaluating it
and g reaers reject any push back without evaluating it, it works both ways. the middle ground is like i repeat, use your AI tools to extract the signal from noise (OP's original point about how AI works)
firefox, brave, whatever all have AI integration built in, you literally open the sidebar and can skip reading any thread and tell the bot directly to just pick out highlights. my rambles are hard for human readers, sure i repeat that's true, but ALL free tier AI tools are able to clearly extract what I'm repeating. Write a geniune reply and I get the same push back with the drawback of being told I've used AI to ghost write the post half the time or people don't reply since they assume its an AI writing it but my writing, being inherently BAD is proof that I am human because why else would you waste your effort to reply to me and to correct me? it means that despite my writing being bad at being good english, it's good at proving that my message is human written and that's better than writing good and being mistaken for a bot.

>>105759778
>If you get lazy and use the LLM all the time you slowly turn into a jeet.
I am already a jeet

>>105759770 (OP)
Main advantage I've seen is that you can get a straight, to the point answer from an LLM. Try asking a question on the Linux forum and you have to wade through dozens of "it works on my machine" posts, demands that you explain why you want to do whatever you're asking about, several people outright calling you a stupid moron for not being able to do something even their blind grandmother can do, several red herring answers, and if you're lucky, maybe one partial answer that sort of works.
The LLM sometimes will hallucinate an incorrect answer that uses something that doesn't actually exist but that's quickly resolved. On a forum, it's another week or two, or maybe months, or maybe forever, waiting for a response as to why a "solution" didn't work.

>>105766112
So you actually admit that it's not really an intentional choice about effort balancing, but just all made up justification in post to hide the fact that you are incapable of writing well?

>>105766352

>>105766352
>but [...] ?
Since you are still here, let's effortpost. I expect a reply, otherwise you undermine your own argument.

Right now, I have about 50 threads hidden in the catalog. These were hidden manually, without filters. At one point yesterday, I had 140 hidden, but threads come and go. They die, new ones are made, and the numbe shifts.

I keep seeing threads like:
>aicg (slop)
>rust noise
>ddg shilling
>politics
>linux flamewars
>eceleb
>obvious shilling
>bideo gaymes
>tech support that should be on wsr
These are not what we'd call "technology". So I make a point to bump actual tech threads, like this one, even if buried on page 10. This thread had two replies: one on-topic, one a last-ditch bump. I added a geniune reply. It was poor written, but it was a real attempt to engage. Now look: 30 replies later and the thread's alive, pushing out the real garbage.

If I had written a clean, polished post like:
>LLMs made me realize Google was shallow, so I started reading primary sources
no one would have replied. it's not bait, it's just boring. My original post was messy, but still earnest. The follow-ups stayed on-topic: "What do you use AI for?"

Even those could've been tidied into:
>LLMs help me extract highlights from tech threads despite the noise
but again, no one would have replied. Why? because geniune, well-phrased attempts get ignored here. People only engage when they think they can call someone (me) a retard. You replied because despite my low effort, you noticed it was human authored unlike the other low-effort threads and posts, because no one replies to hecklers. They're unreachable.

So it is about effort balancing and it works! My posts despite what you feel about their appearance, still carry some earnest messaging that you would had otherwise ignored had done what you suggest. This post I edited better, not well but better. I know the effort is unlikely to be matched but my expectations are also low and so the effort balances.

>>105766580
This is so obviously written/edited by an AI that it's funny you thought you'd pass muster. Are YOU capable of writing well? The answer is "no".

>>105759778
They hated him because he told the truth

>>105766598
>This post I edited better, not well but better. I know the effort is unlikely to be matched but my expectations are also low and so the effort balances.
as expected, had it been written by an AI it would not had messed up writing
>number
as
>numbe
and wouldn't had missed the "I" in
>had "I" done
i proofread once, and send, i read again because I'm aware my own writing is poor without revision, not because I cannot write at all. Good writing is rare in any context (see the Orwell post again, it's a short essay that conveys what you're dancing around)

>>105766631
>I'm aware my own writing is poor without revision
And apparently it's poor with revision too. We're back at my point that you can't write, so you make up all this stuff in post about why writing well is actually pointless. It's cope.

>>105759778
I tried this approach for awhile, but honestly the Google searching, Stack Overflow comment hunting, and poorly written readme’s were so obviously less efficient that I just replaced all my Google searches with ChatGPT instead.

For example, I’ve been working with Azure for the last three years. I’ve found their guides are usually super basic and at times even incomplete. They’re poorly worded and lack depth, so anything in trying to accomplish that isn’t boilerplate is extremely difficult to find information for on their official guides (usually, but not always). However, ChatGPT always points me in the write direction with one well-worded prompt.

It’s just too efficient; I can’t go back to struggling to understand some shit code posted on SO.

>>105766692
>We're back at my point that
>you actually admit that it's [...] really an intentional choice about effort balancing
all writing needs revision, let's edit your post
>And apparently
dont need this if the point is to say my writing is poor, the too at the end is also unnecessary. you can skip the part where you go "we're back at my point that", you can just be assertative like this
>it's poor with revision. you can't write. it's cope
but the problem is that you can just carry it all even better by just saying i'm retarded, that's the problem, there's no reason to write much at all here, you dont have to dance around like you have to do on the orange site HN, you can just say i'm retarded and you get the same output and that means even your writing is bad when the simple and direct is available (ergo read the orwell essay i mention, again it's really good)

>>105766580
My guy, you sound retarded. You’re trying so hard to sound smart that you just come off like a sperg.

>>105766752
>>105766747
see here

>>105759770 (OP)
>I think I've learned more in the last two years than I did in the first 10 without them. Probably because I don't feel limited by skill anymore
Yeah, reading is extremely hard for some people, mostly retards.

>>105760222
>>105760486
>>105760638
Braindead zoomies detected