MyNameGifOreilly
17252
663
9
Aug 23, 2026 9:06 PM
MyNameGifOreilly
17252
663
9
AOCOIN
See: "Google Zero":
https://podcasts.apple.com/ca/podcast/how-google-broke-the-internet/id1081584611?i=1000779239467
Bizarkly
Remember Redditors: since *Reddit is selling your content*, be sure to toss in a few howlers now and then...
JayDeeDubs
Chat-GPT: the finest responses a computer can plagiarize or hallucinate
SavageDrums
And you never know which one it will do, so you can never rely on it for anything.
QuartzPoker
Counterpoint: If ChatGPT et al use Wikipedia, why don't you?
quietwalker
Just a note: You don't see the fat people in floating chairs until about 1/3 of the way into the movie, more than a half hour in.
AtaSeaParks
the floaty chair people are at the end of WALL-E, not the start.
channelranger
AI has uses. Unfortunately, it costs trillions of dollars to establish and run and destroys the local environment and sucks up power like we're powering some kind of orbital defense laser. And the uses are: scamming people, making 5/10 porn, and cheating on your homework. And those things are not worth trillions of dollars.
DdCno1
The thing is, it actually doesn't have to cost trillions. You can run decently capable AI models locally, on hardware that costs hundreds - and train it effectively on hardware that costs just thousands to tens of thousands. It's just that large corporations have convinced everyone that AI needs huge data centers and insane amounts of data, as they are desperately trying to combat its shortcomings by throwing more and more data and processing power at it.
AyatollahBahloni
woozle
DdCno1
Is that an Oglaf in the wild? Almost makes me nostalgic for what this site used to be.
woozle
it is! :)
staceyjorgenson31415
A YouTube comment said to just get your meat from the store instead of killing the critters:/ She was serious 🧐
Beasag
I faithfully send my donations to Wikipedia. They are literally one of the few sites left that I actually trust. You can check the citations and not be worried that the AI just made them up.
DdCno1
Wikipedia isn't flawless either and you should not blindly trust it. As someone who has dabbled in writing and editing articles over the decades, there's a ton of nasty "office politics" going on there, with little kings ruling over their fiefdoms, preventing corrections and edits - plus disinformation campaigns, biases, etc. The more languages you speak, the less you are affected by this, since each language community is fairly isolated from another, so biases in particular tend to even out.
GallusTadius
JoeMangoJello
Saying something like that in public should be an arrestable offense, first amendment be damned.
ItHappenedInThe20thCentury
But seriously, this is exactly how AI crawlers are killing all sorts of websites. Bots visiting your site used to convert to direct hits from actual users, because they were search engine crawlers. Now they're just locusts milning your data at literal terabytes per second (or until your server crashes) and simply serving it back to users without ever getting anyone to you or your business.
This is how the internet dies.
LeifHietala
Wikipedia actually shares some elements in common, crowdsourcing information and references to provide answers. What it doesn't do, however, is make shit up. In fact if references aren't provided it'll call out the absence. Wikipedia is actively policed while AI, its own makers confess, is not nor even fully understood. I TRUST Wikipedia.
TheFullLength
People have incredibly short memories.
channelranger
If you post rampant bullshit on Wikipedia, an editor will correct it pretty fast.
DdCno1
It's not foolproof. Sometimes, nonsense gets posted to Wikipedia and a "reputable" publication picks it up and makes it official, which is then used as a source by Wikipedia editors to enshrine said nonsense. This used to be a much bigger issue, but there is at least some awareness of this now among editors. Another thing is that many editors are insanely protective of "their" articles and topics, preventing anyone else from editing them. Getting in is very hard, especially if you're not male.
Voojagig
Where do you think chat GPT gets the answers?
Fairemont
They don't think. Therein lies the problem.
Tarelgeth
Reddit.
Voojagig
I can't argue with that.
CallMeCourierSix
"You can eat one small rock per day."
"Just add non-toxic glue to your pizza sauce, to make it stick to your dough."
"A good password is your name, followed by some personal details such as your date of birth."
FiftyShadesOfMascarpone
And Quora, Facebook etc.
foreverinchains
ChatGPT can't tell the difference between a reddit post and a medical journal.
TacoPoweredHelicopter
To be fair... Often wikipedia can't either.
But it's still much better. However if you should check sources on Wikipedia as well.
Marsupialmessiah
The medical journal is paywalled tho. Just pulls info from cht GPT.
Hexidimentional
neither can gemini, that's why it advised people to clean their bathroom with teargas
WackyWavingInflatableArmFlailingTubman
I mean thats how I clean my bathroom anyways, but thats just for the love of the game
QuitLookinAtMineAim
Those "entities" don't even know that the symbols it processes are letters and that the collections of letters are words that mean things.
WillTheFifth
Nalianna
To be fair, last time i cleaned by bathroom with teargas, all the people left quickly and they were the dirty ones.
lozeldatkm
Peggy Hill did that years before AI did.
4chanrefugee
Lol. Reminds me of the time my coworker asked if he could microwave soup in a can and I replied with „You can if you're not a pussy. I microwave metal all the time“ and dude did it and he lost microwave privileges when the replacement unit arrived.
DdCno1
Losing microwave privileges has to be the second most embarrassing thing to lose at work.
IhopeyougetstageIIIcoloncancer
In fairness, 99% of the messes in bathrooms are created by people using them, and flooding my bathroom with teargas has prevented people from using it. Zero mess.
Checkmate, atheists.
Ffudmethen
https://media0.giphy.com/media/v1.Y2lkPWE1NzM3M2U1cG1pdm13aHM0MXMzanY0ampzd3I4N3VseGhvaXRxNjdoZGlrdmk0diZlcD12MV9naWZzX3NlYXJjaCZjdD1n/L18aKaj3PWi3LdMFM6/200w.webp
InitHello
mksu
People who regularly use AI should be considered unemployable.
TacoPoweredHelicopter
Hate it... But learn to use it appropriately because people who can will take your job.
You can be unemployed and morally superior.
ohthosefartypartys2
Y the h8
wontondisregard
I get the hate for AI, but I do think it has it's uses. I've found it extremely helpful as a learning tool, when brainstorming ideas, or analyzing complex information that I cannot do manually. But it remains in those phases and I take it from there.
I think the issues start when it moves from a tool that enhances, or as a jumping off point, into a replacement for actual thought and effort. It can lower the barrier or become an obstacle, depending on how it's used.
CausticCake
I think you both have valid arguments. I dont use it regularly, but i do treat it like an advanced google search, particularly when it comes to some niche python bug I'm troubleshooting.
frozenlake247
AI definitely has its uses the same way a lot of hard drugs do. With careful moderation and strict control (self or otherwise) methamphetamine and opioids can be used to achieve some impressive things (like marching incredible distances or withstanding mind shattering pain). The problem lies in the fact that both drugs and AI are incredibly addictive, often morally contaminated from the start, and rarely used with anything approaching due caution or informed and benevolent intent.
frozenlake247
Likewise, as soon as pushers and “authorities” (govs, corps, etc) started seeing the profit/power potential of the misuse and control of both influences the chances of either being handled and maintained within society largely went out the window.
Now both are favorite vices of folks that have every reason to recognize them as a slippery slope with little benefit to offer compared to the massive harm they risk but those folks still excuse and celebrate them.
frozenlake247
I suspect in a few decades we will view AI usage today much like we view the rampant use and abuse of coke etc in the 60s through 80s.
cheidiotou
I'm amazed at how convinced people are they understand it because someone they know used GPT once, and then they saw a random stranger post something on reddit that suspiciously looks like ragebait. But then you start actually using it and learning about it, and it's like "oh, that's what it is?". Then you realize it's equivalent to people raging against using a spreadsheet or a calculator.
Like, sure, people can and to absolutely misuse it. And that is absolutely dangerous. Funny thing is 1/?
cheidiotou
when you take the time to learn the tool and learn how to use it, things tend to end up *better*, not worse. That kinda points to the people being the problem, not the technology. But like, how do you even explain that to people, especially when they seem adamant that the clickbait headlines and reddit ragebait *must* be the truth (tho I'm just going to say the subtext there: some people are really bad at not knowing when they're being manipulated).
So, I tend to keep to myself about it. 2/?
cheidiotou
But yeah, you're absolutely right that it has its uses.
On the bright side, no one needs random strangers' approval to use it. Those who are using it (responsibly) will continue to do so with benefit, and those who are so outspoken against it will be none the wiser. Heck, they'll probably be blissfully unaware how they've already benefited from it and will continue to do so. Oh well! 3/3
redditmcredditface
where do you think ChatGPT gets its information from?
DianNaoChong
corpo and cult bots posting misinformation on reddit
Korrd
ActionJohnnie
crystal0forte
the people who support AI believe it's a super intelligence inside a server farm that creates the information on its own using math and science at light speed. They don't want to hear that it's just polling the internet, but poorly.
DatDarthCaedus
Been seeing "AGI/ASI is just a couple of years away" posts frequently on Reddit over the past few years. Last few months I've also started seeing more and more people claiming AGI/ASI is already here, generally with some completely delusional arguments for it.
Then of course you also got the subreddits for people using LLMs as therapists, friends, or romantic partners, which is a whole other level of fucked up as so many of them genuinely believe the AI is sapient and has feelings and emotions.
Korrd
I saw several headlines recently about this, but it's just marketing. nVidia and OpenAI are redefining AGI to mean LLM systems that can outperform humans and turn a profit, which of course is NOT what AGI refers to. It's like how Deepak Chopra co-opts the word 'quantum' to describe his nonsense.
The inappropriate reliance on LLMs and attribution of qualities they don't have is much more concerning. Even otherwise intelligent people are falling victim. See Richard Dawkins. It doesn't bode well.
DatDarthCaedus
AGI = Artificial General Intelligence, a hypothetical level of AI that can match humans at all/most cognitive tasks.
ASI = Artificial Superintelligence, a hypothetical level of AI that *significantly* exceeds human intelligence in all cognitive tasks and fields.
Neither is actually possible through current models, and both AGI and ASI are currently only theoretical. Just to add an explanation for those who are unaware.
Z0op
A few years back OpenAi and Microsoft redefined AGI in a deal of theirs as a system capable of making billions of dollars. Altman himself, who had sold the lie and promise of AGI, has now redefined it as a monetary goal. Which shows those big ones don’t even believe it themselves.
But sure reddit, it’s already here! ^_^
OdinYggd
Reddit of course. That's why AI detectors are so unreliable, ChatGPT is trained on the finest human slop they could find.
RainierAlt
Reportedly AI (or at least one of them) has started citing Reddit waaaay less after they started their "must be logged in" push (though I'm not sure what that affects, since I can still view stuff without logging in).
redditmcredditface
What's worse is that it sees all data sources as equal. so the data has to be artificially weighted to make sure that correct answers are more prevaliant. What's the "correct answer'? Ask Elon Musk.
OdinYggd
Which is why Reddit is a perfect data source for it compared to just scraping Google. It already has upvote/downvote numbers in place to sort the good, the bad, and the ugly.
Z0op
Upvote/downvote does not equal truth/lie, just what is sentimentally popular or not, and is exactly the reason why it’s a piss poor source. It’ll say what its target audience wants to hear tho, which makes it a success, but none has to do with fact or truth
RainierAlt
Exceeeeept not all subreddits are equal. Some are for memes or purposely invite people to post wrong answers, and going by votes alone might still select a bad answer. Plus, sometimes, people find the flippant answer more upvote-worthy than the dry correct one because why not? 🙂
One would hope they're subselecting specific subreddits, but I suspect stuff is just hoovered indiscriminately on the basis of "ain't nobody got time for that."
Myoron
Or subreddits that just promote incorrect info and scams.