This. LLMs aren't going to become intelligent, but they can fake it well enough that morons will think they are. Those are the people we need to worry about.
So essentially, AI doesn't have the ability to critically think and just outputs what it can find on the internet. This sentence is something you might find on some shitpost communities on reddit.
Even worse than that: It's just tokens. They don't even know what language they're using or even that they're putting words together. It's just, "this arbitrary thing is mathematically based on a limited model most likely to proceed that arbitrary thing".
and yet what? Yet idiots think it has intelligence? And yet it usually produces understandable results? And yet it will still replace enough jobs to fuck over the economy for millions upon millions of people?
mathematics shaped into "conversation" where the bottom line is "words that a human will consider conversationally consistent, when reckoned against the previous words" -
This is the Google search overview AI. It just reads the top results and summarizes them together. You don't directly prompt it, it's already prompted to just do that. The problem with that arrangement, as demonstrated here, is it will confidently and non-critically summarize parody, idiotic rambling, intentional misinformation and any other sort of nonsense that the search algorithm pulls up.
Huh, never tried it, and never done it by accident that I noticed. Makes sense though. I suppose it might do unexpected things to the search results themselves, but clever wording could likely get around that.
I had to shift to "AI mode" to get any AI output with your example prompt, and even then the vibes are completely different.
I'm sure there's a way to engineer the query to guide the AI overview to a desired response (or it's just entirely faked) but that was less effective than I anticipated
I feel like I've gone mad any time I read an AI output.
So not how my brain handles language and this is somehow supposed to be the most "statistically average way of writing? Even as a former tumbler native this is just... Cotton candy nothing.
That is absolutely correct, and the deeper problem is it typically "writes" in a tone that fools people into feeling like it can do so.
just outputs what it can find on the internet
That is what this one, the Google search overview, is set up to do. Other LLMs don't, or may only do so when prompted to.
What they all do is analyze the patterns of all the words that have already been input, such as the initial system prompt, the user's prompt, any sources it's referencing and any replies it's already generated, and then it uses that to predict what words ought to come next. It uses an enormous web of patterns pulled from all of its initial training data to make that guess. Patterns like a question is usually followed by an answer on the same topic, a sentence's subject is usually followed by a predicate, writing tone usually doesn't change, etc. All the rules of grammar it follows and all the facts it "knows" are just patterns of meaningless symbols to it. Essentially, no analysis, logic, or comprehension of any kind is part of its process.
Be it from an actual teenager or possibly it's a very specific reddit post or something that is basically being copy pasted. Like that one with the rocks.
Not much to say. Basically there was a reddit post about eating rocks or some shit, one of the responses was basically you should eat 3 to 6 rocks a week. An AI scraped the data and gave it as an answer verbatim.
Fucken hell didn't realize it was that bad. I just remember folks talking about it and seeing a couple screenshots didn't realize they had that piss poor of a filter even then, but then again what should I expect from silicon valley?
Oh my god, from now on I will willfully misinterpret every AI requirement I get to "outsource it to India". Still crap, but loads better than the alternative.
Maybe not, but it seems like this career has died. Haven't seen him doing anything in the last 5 years, and I don't think there's been an album published under his name in more than 15 years.
The man might not be dead, but it sure seems like his career is.
If that diet ever only applied to the boobs and butts, guys would be buying them for you everyday. Sadly, the diet blows you up all over to the point your flopping over your undies and slapping your knees with your navels
65 Comments
SantasMagicalComfort@piefed.world · 120 pts · 274d
Why the fuck did you cut out the information on sleeper fatties?
xc2215x@lemmy.world · 87 pts · 275d
A sleeper fatty ? What ?
LaserTurboShark69@sh.itjust.works · 54 pts · 275d
She got a sleeper fatty
(Secretly has a fat ass)
AllNewTypeFace@leminal.space · 35 pts · 275d
That’s like some sort of gluteal Tardis, thiccer on the inside than the outside?
RickyRigatoni@retrolemmy.com · 16 pts · 274d
If you can keep the fatness of the ass a secret it isn't fat enough.
LaserTurboShark69@sh.itjust.works · 12 pts · 274d
You could do a PHD on that thesis
BakerBagel@midwest.social · 21 pts · 275d
Sleeper fatty is what i am gonna start calling my morning wood
AstaKask@lemmy.cafe · 70 pts · 275d
Don't fear AI, fear the people who think an LLM can be intelligent, for they are as dumb as they are gullible.
UnspecificGravity@piefed.social · 26 pts · 274d
This. LLMs aren't going to become intelligent, but they can fake it well enough that morons will think they are. Those are the people we need to worry about.
TexasDrunk@lemmy.world · 8 pts · 274d
Many vulnerable people, morons, and heavy ketamine users already think it.
OpenStars@piefed.social · 7 pts · 274d
What if, hypothetically speaking, this describes your boss?
UnspecificGravity@piefed.social · 10 pts · 274d
This describes most people's bosses.
lightnsfw@reddthat.com · 2 pts · 273d
It's not like morons embracing the absolute worst trends in technology and dragging us all down with them is a new occurrence.
UnspecificGravity@piefed.social · 3 pts · 273d
Reminds me of that five year period where every movie had to be shot with a 3d release in mind for no reason.
julietOscarEcho@sh.itjust.works · 3 pts · 274d
And by that we mean both that the people are dumb/gullible and that the LLMs are dumb/gullible.
Una@europe.pub · 40 pts · 275d
So essentially, AI doesn't have the ability to critically think and just outputs what it can find on the internet. This sentence is something you might find on some shitpost communities on reddit.
cv_octavio@piefed.ca · 15 pts · 274d
MotoAsh@piefed.social · 4 pts · 274d
Even worse than that: It's just tokens. They don't even know what language they're using or even that they're putting words together. It's just, "this arbitrary thing is mathematically based on a limited model most likely to proceed that arbitrary thing".
zarkanian@sh.itjust.works · 1 pts · 273d
And yet
MotoAsh@piefed.social · 2 pts · 273d
and yet what? Yet idiots think it has intelligence? And yet it usually produces understandable results? And yet it will still replace enough jobs to fuck over the economy for millions upon millions of people?
julietOscarEcho@sh.itjust.works · 2 pts · 274d
Me to, transformers, me too.
village604@adultswim.fan · 8 pts · 274d
The thing to remember about these kinds of posts is that they never show the prompt. There's a good chance they instructed the AI to respond this way.
EpeeGnome@feddit.online · 12 pts · 274d
This is the Google search overview AI. It just reads the top results and summarizes them together. You don't directly prompt it, it's already prompted to just do that. The problem with that arrangement, as demonstrated here, is it will confidently and non-critically summarize parody, idiotic rambling, intentional misinformation and any other sort of nonsense that the search algorithm pulls up.
village604@adultswim.fan · 6 pts · 274d
You can absolutely prompt the Google AI overview with your search query.
EpeeGnome@feddit.online · 2 pts · 274d
Huh, never tried it, and never done it by accident that I noticed. Makes sense though. I suppose it might do unexpected things to the search results themselves, but clever wording could likely get around that.
hobovision@mander.xyz · 8 pts · 274d
I'm pretty sure the thing to remember is that these posts can easily be faked.
pelespirit@sh.itjust.works · 3 pts · 274d
How could you prompt the AI to say this by the way it responded?
Deceptichum@quokk.au · 3 pts · 274d
Trainguyrom@reddthat.com · 8 pts · 274d
I had to shift to "AI mode" to get any AI output with your example prompt, and even then the vibes are completely different.
I'm sure there's a way to engineer the query to guide the AI overview to a desired response (or it's just entirely faked) but that was less effective than I anticipated
Krauerking@lemy.lol · 3 pts · 273d
I feel like I've gone mad any time I read an AI output.
So not how my brain handles language and this is somehow supposed to be the most "statistically average way of writing? Even as a former tumbler native this is just... Cotton candy nothing.
village604@adultswim.fan · 2 pts · 274d
You can still prompt the search AI with your query. I do it all the time for basic troubleshooting or command building.
EpeeGnome@feddit.online · 3 pts · 274d
That is absolutely correct, and the deeper problem is it typically "writes" in a tone that fools people into feeling like it can do so.
That is what this one, the Google search overview, is set up to do. Other LLMs don't, or may only do so when prompted to.
What they all do is analyze the patterns of all the words that have already been input, such as the initial system prompt, the user's prompt, any sources it's referencing and any replies it's already generated, and then it uses that to predict what words ought to come next. It uses an enormous web of patterns pulled from all of its initial training data to make that guess. Patterns like a question is usually followed by an answer on the same topic, a sentence's subject is usually followed by a predicate, writing tone usually doesn't change, etc. All the rules of grammar it follows and all the facts it "knows" are just patterns of meaningless symbols to it. Essentially, no analysis, logic, or comprehension of any kind is part of its process.
zarkanian@sh.itjust.works · 2 pts · 273d
And nobody knows how to get from that to actual thought. And if they tell you that they do, they're either deluded or lying to you.
A lot of tech bros have been getting high on their own supply and think that we're right around the corner from a Skynet situation.
Zexks@lemmy.world · 1 pts · 275d
crunchy@lemmy.dbzer0.com · 35 pts · 275d
Why does this "AI" read like it was written by an Indian teenager?
village604@adultswim.fan · 37 pts · 274d
It literally might be.
vaultdweller013@sh.itjust.works · 5 pts · 274d
Be it from an actual teenager or possibly it's a very specific reddit post or something that is basically being copy pasted. Like that one with the rocks.
xorollo@leminal.space · 3 pts · 274d
Tell me the story about the rocks please
vaultdweller013@sh.itjust.works · 9 pts · 274d
Not much to say. Basically there was a reddit post about eating rocks or some shit, one of the responses was basically you should eat 3 to 6 rocks a week. An AI scraped the data and gave it as an answer verbatim.
hovercat@lemmy.blahaj.zone · 3 pts · 273d
It's worse because the rocks thing was literally from a fucking The Onion article. Adding glue to pizza to get the cheese to stick was the Reddit post
vaultdweller013@sh.itjust.works · 2 pts · 273d
Fucken hell didn't realize it was that bad. I just remember folks talking about it and seeing a couple screenshots didn't realize they had that piss poor of a filter even then, but then again what should I expect from silicon valley?
zarkanian@sh.itjust.works · 3 pts · 273d
AI = Actually Indian.
FatVegan@leminal.space · 24 pts · 274d
So AI is actually short for An Indian...
Whostosay@sh.itjust.works · 6 pts · 274d
This is true in many cases lol
markstos@lemmy.world · 3 pts · 274d
Years ago I used a voicemail to text service that worked like that— It was powered by human transcribers.
Randelung@lemmy.world · 2 pts · 274d
Oh my god, from now on I will willfully misinterpret every AI requirement I get to "outsource it to India". Still crap, but loads better than the alternative.
Also better for the environment and the economy.
pelespirit@sh.itjust.works · 23 pts · 274d
'700 Indian engineers posed as AI': The London startup that took Microsoft for a ride
Bosht@lemmy.world · 28 pts · 274d
What the fuck is a sleepr fattie?!? I MUST KNOW
SkunkWorkz@lemmy.world · 6 pts · 273d
It’s like a sleeper cell. If the sleeper fatty receives a subliminal message they hijack an airplane and crash into the nearest skyscraper
lastunusedusername2@sh.itjust.works · 6 pts · 274d
I have a truly marvelous explanation but my sleeper fatty is too fat to leave room for it...
OpenStars@piefed.social · 5 pts · 274d
Too ba...
bigfondue@lemmy.world · 3 pts · 273d
MystikIncarnate@lemmy.ca · 20 pts · 274d
I like big butts and I cannot lie!
Bonesince1997@lemmy.world · 14 pts · 274d
May I offer you 6-8 donuts?
HackerJoe@sh.itjust.works · 4 pts · 274d
DancingBear@midwest.social · 2 pts · 273d
agamemnonymous@sh.itjust.works · 1 pts · 273d
I like big butts and I cannot tell the truth!
You must best our challenge if you wish to pass, wayward traveler.
MystikIncarnate@lemmy.ca · 1 pts · 272d
Sir mix-a-lot didn't die so I can experience such oppression!
agamemnonymous@sh.itjust.works · 3 pts · 272d
Sir Mix-a-Lot isn't dead
MystikIncarnate@lemmy.ca · 2 pts · 272d
Maybe not, but it seems like this career has died. Haven't seen him doing anything in the last 5 years, and I don't think there's been an album published under his name in more than 15 years.
The man might not be dead, but it sure seems like his career is.
I hope he's enjoying retirement.
agamemnonymous@sh.itjust.works · 1 pts · 272d
Oh, I was setting you up for "Like I said, Sir Mix-a-Lot didn't die"
MystikIncarnate@lemmy.ca · 1 pts · 272d
Right. Sorry, I've been dense the last few days.
Have a good one, kind stranger.
DancingBear@midwest.social · 1 pts · 269d
Kolanaki@pawb.social · 10 pts · 274d
Is "sleeper fatty" just another way of saying "morning wood?" 🤔
OpenStars@piefed.social · 2 pts · 274d
There is only one way to know for sure!
(Which I for one have no idea as to what that is)
DancingBear@midwest.social · 3 pts · 269d
OpenStars@piefed.social · 0 pts · 269d
OpenStars@piefed.social · 9 pts · 275d
You certainly will not regret eating 6-8 donuts a day, thereby becoming phat and wealthy with zero negative consequences whatsoever.
TheLeadenSea@sh.itjust.works · 6 pts · 275d
Ah, it's that time of year.
Sam_Bass@lemmy.world · 2 pts · 273d
If that diet ever only applied to the boobs and butts, guys would be buying them for you everyday. Sadly, the diet blows you up all over to the point your flopping over your undies and slapping your knees with your navels
FelixCress@lemmy.world · 2 pts · 275d
I support this message.
smh@slrpnk.net · 1 pts · 274d
MNByChoice@midwest.social · 1 pts · 275d
The "Fear of AI" is advertising for AI.
Eat those donuts.