You must log in or register to comment.
In my time we didn’t paste LLM-generated code we barely understand and hoped it compiled, let alone work. We pasted code from stack overflow we barely understood and hoped it compiled and let alone work, as god intended.
God has no hand in programming. He’s just as confused as us.
Oh there’s definitely some elder gods involved with programming when I do it.
User Feedback, the Crawling Chaos, the Haunter of the Dark… I feel its tendrils of madness reaching for my mind even now. I am not ready for this. Ph’nglui mglw’nafh caffeine R’lyeh wgah’nagl fhtagn! Iä! Iä!
Ah, I see you’re a fellow html regex parsing enthusiast.
I thought mine was funny because I had not seen that, and I am humbled. Damn. Fukken saved.
I am a better programmer than God, peace be upon Him. This implementation of
kneesis Exhibit 1.ok but real talk, knees are genuinely one of the most marvellous pieces of biomechanical engineering. They can withstand decades of constant movement, can allow extension (with a lot of force) even when bent 180°, can withstand - and move - hundreds of kg per knee (with enough practice) periodically also for decades, and can comfortably remain with your entire body weight resting on them at any angle from 0 to 180° for any length of time. It’s amazing that everyone doesn’t have constant knee pain or have their knees simply fail altogether.
As a representative of those who have had a constant knee pain for over a decade: I’m slightly less thrilled about the design.
Now we’re still pasting code from stack overflow we don’t understand, we’re just getting it from an LLM
At least now I don’t have to deal with the rudeners.
I do like the fact that when I ask it a question it actually gives me the answer, and doesn’t tell me to refactor my entire code because apparently I’m a bad person.
Weirdly enough, when someone asked an LLM for an OpenGL grievance for me, first it just recommended to use GLFW or SDL for the task, both of which I didn’t want to use (GLFW controller handling🤮), after that they got an answer that just bugged out on most WM.
Yes, but did the LLM get it from the answers or the questions?..
Who knows, and furthermore, who cares?
You’re young. Back in my day, we bought a book called “Advanced Algorithms for C vol. 3”, and we manually typed the code from it if it didn’t come with a CD.
I’m too young for that, but I got a piece of that experience when I bought a physical programming book as a reward from Kickstarter.
Some of the code lines were too long to fit the page and were cut off which added another fun element (though it was pretty rare).
I think the same people who run stackoverflow must run a ton of subs on reddit.
“Your post was removed because it uses “the” too much and doesn’t contain enough w’s and because the moon is in Pisces and it’s Saturday. If you think this was done in error please message the moderators.”
messages moderator about it, banned from subreddit for no reason given. Or at least that is how i imagine how it would go
moderator shoots anyone as soon as they knock on their front door through a custom slot
INTERCAL’s PLEASE Politesse Checking
That’s hilarious. I do hope it gets evaluated at run time. That way you could have a program that works most of the time but if some rare circumstance caused it to execute commands in a sequence where the correct level of politeness was not maintained it would get the hump and crash
They’re the same people posting articles saying how bad AI is while everybody else has fun using it
Nobody cares if you are “playing” with AI. The problem is companies like OpenAI stealing people’s work for training, and stealing their business using that stolen work for profit. Fair use stops at commercial interest and AI is nothing but commercial interest. It’s a layer cake of theft and unfair competition.
Don’t care, of all the problems in the world. You’re all making shit up to be mad about. Everybody gave their stuff up with or without AI. People were always scraping data and allowing it to be locked awa. Information brokers are not new. They have always sold our data. Nothing has changed.
I’m not bending over backwards to give a shit for people who have slept to the point that we arrived here with these issues, and now that we finally have something cool they’re up in arms because they think their stupid Garfield knockoff comic is the core competent to making AI powerfully profitable. I’d all for removing every content creator from the Internet. Scrub the whole fucking thing. They’re parasites
There is no openAI without content creators. It’s a derivative tool, so if some new tool/system comes out it’s not gonna know jack shit without the training data.
Now if your beef is with the content platforms, I completely understand that, but saying content creators should be removed, seems a bit myopic
I fully believe we lost the Internet due to content creators. Once PewDiePie made that first million it was all over. That moment is the point in our timeline where we veered into a very dark timeline instead of the good one. The other point was the Hanging Chad
Barking up the wrong tree?
IMO, this would be more ironic if the post was closed automatically by a bot. But that’s not the vibe I’m getting from this.
Thanks Cloudflare for giving me a moment of reflection on why the fuck I am heading to Stack Overflow so I can close the tab before I get there.
CF: We defended your website from 69,420 bots today!
The 65,000 users: 👁️L👁️
its funny. when its anubis, the common opinion is rightly fuck ai, but when its cloudflare, then it is somehow fuck the website.
what a weird world we live in.
Anubis does its thing, shows me cute art, then leaves without elaborating. It’s a mostly non-intrusive, individual/community effort to protect people against big tech and abusive scrapers. I usually see it in open source community websites that were getting hammered by LLM scrapers.
Cloudflare’s is a corporate solution from the company that man-in-the-middles half the internet and makes me click shit every fucking time. I see it whenever I make the mistake of following a stackoverflow link.
You’re goddamn right my reaction is accordingly different.
Cloudflare’s is a corporate solution from the company that man-in-the-middles half the internet and makes me click shit every fucking time. I see it whenever I make the mistake of following a stackoverflow link.
on SO sites all the difference is a single click. and you have to allow scripts and cookies for both, so no difference there.
Mods be thinking that if they dig SO’s grave deep enough it will emerge on the other side of the world.
I wouldn’t call stackoverflow reliable. It is only partly reliable, if you are lucky.
Thread closed because that’s a stupid question and you should feel bad about yourself.
Not necessarily about stack overflow. But i just got myself in a situation where the first search result I found for a problem was clearly AI generated. And the solution it provided was not at all technically possible. The AI decline is really terrible…
That said, does anyone know of an extension or block list for those terrible AI slob websites? Or a way to filter it from duckduckgo?
well yeah they went all in on ai.
I use SO all the time and I truly had no idea… You mean a lot of answers are submitted by users who used AI?
other way around. they pivoted to offering enterprise solutions based on ai interpretations of their database to business customers. only they were too slow, since everyone had already scraped them.
Good riddance. Whenever I search for a programming question I’d always hope for a) an official documentation page or, failing that, b) a page on a dedicated forum for the tool that I was using that covered the problem. I’d only ever click on SO links if I had no other choice.
And, of course, I’d never search for a problem on SO itself.
SO used to be good, but they have this problem right down in their core concept that makes sure the content gets outdated fast.
And that’s the concept that every question can only be asked once.
That makes sure that everything gets outdated as soon as possible.
- Q: Can X be done in framework Y? (asked in 2012)
- A: No.
Now it’s 13 years later, and framework Y can do X since 5 years, but you can’t ask again, because your question will get closed as a duplicate to the outdated one from 2012. And since every time someone asked this question again in the last 13 years the question just got closed, google will just link you back to the question from 2012 claiming that framework Y can’t do X.
I hate that so many projects are moving from public support forums to fucking Discord channels. God forbid a tech project be expected to maintain a public indexable forum and website. You can’t search it unless you join the channel, it’s not well organized at all, and the invite link probably expired 3 months ago. Fuck you if you didn’t join while it still worked I guess.
I almost always prefer SO answers because there was chance someone had the same issue I was seeing. Documentation only shows how things should work and dedicated forums are very hit or miss.
SO used to be really good in the past, but these days when I’m looking for an answer to a problem, I only unanswered closed questions.
That’s the stupid SO concept that questions can only be asked once. It makes sure that only the question from 2012 with all its outdated answers is available while everyone who asks for an update will get their question closed.
That’s why SO was great in 2012 but its rules made sure it got outdated fast.
this is a crap question, glad it was closed
