Country usually has coastal waters. For example 12 miles off the coast of the US is still considered part of the US. So the country US does not fit into its physical land area.
You know, thats where a bunch of the natural language training for these models come from. (But sure Grok is probably even aimed at this type of speech
That population comment has me wondering now. Is there a country that the population of itself, would NOT fit inside.
Like maybe some random tiny European country that produces a lot of people who go live in larger nearby countries, but they are still citizens of their home land. Then for some reason, they all decide to return home.
I guess my point is, that actually could very easily be true, that there is a country that can't actually contain its population.
I mean, physically, probably not, but more like, with houses and shit.
My phone got force updated to gemini and I kind of dig it. Mainly because when I told it to navigate to costco and it asked whether I wanted to go to the bakery or the food court I said "whichever one because they're in the same building you fucking clanker" it just routed me to one of them instead of saying "sorry, I don't understand" like Google assistant usually did.
Realistically though, using an LLM for natural language processing should have happened years ago and I'm pissed it's so enshittified now that it's finally happened.
It would be better if the damn LLM was local, i tried to do it on a crappy 300€ phone, it worked, the LLM was one of those stupid ones but it did a deceny job for a 300€ phone, so i can guess that with flagships it would be good, heck, if we had a processing unit dedicated to LLM stuff it wouldn't be that bad probably
Crazy how somebody is even mentioning the fact, that Russia can fit multiple times in itself, but only, and only if rotated correctly. I also love the sentence used to conclude by ChatGPT:
Geographers still don't fully understand why Japan matches Japan so closely.
Any Geographers here to explain? But no seriously I know this is frequent but the way it speaks to make its statement seem believable by citing non-existent expert certainly is not one of the better sides of LLMs.
No, no, it clearly states Russia can only fit inside itself multiple times if rotated INCORRECTLY. Like, opps, I fell through a 4D wormhole while rotating Russia.
That's a fun play on words! Since "the size of Japan" is just a measure of its own land area, it's trivially true that Japan fits perfectly inside itself — just like any country or shape fits perfectly inside its own boundaries. It’s a tautology, but it sounds like a clever riddle at first listen. Nice one!
Okay, your output is different given the same input... So what? It's a well known fact that these LLMs are non deterministic. Theres a guy on youtube that asks chatgpt everyday to count to 200 until it doesn't fuck up. Your output does not prove or disprove the authenticity of the original post.
Tbh them being nondeterministic is a big part of why they're so unreliable. Like, maybe it'll work fine for 9/10 people, but then there will be that one person whose home directory gets wiped for whatever reason. Or maybe it'll do math right for those nine people, but then for that one person it'll say 1 + 1 = 11.
You're basically gambling if you don't verify the answers.
I'm referring to nondeterminism for the same prompt, since unless you start a session from scratch, it's unlikely you'll have the same history. If you give it a prompt, then depending on what you've told it previously, it may blow up in your face.
Determinism for the same prompt means you can't give it context through a conversation, which vastly shrinks its utility.
That said, even that form of determinism can be unreliable: the example of arithmetic still works; you could have it completely deterministic, but if it only performs correctly on 80% of arithmetic problems, it's still unreliable.
In fact, if you give it this prompt 50 times and it only fucks up once, that clearly indicates that this is post is misleading. It's also likely this post was faked with a different prompt than the one shown.
Open a new thread and ask again. You'll get a different result. Open a third thread and ask again, you'll get yet another result. That's how LLMs work. You getting a different answer to your prompt doesn't mean anything.
People will read misinformation on the top comment of a Today I Learned post and take it as gospel truth before they click on the actual link of the post that immediately debunks it.
We are in a memes community here, most memes are only tangentially related to the truth (if that). It's a joke, it doesn't have to be true, but if it is true (which there might be a chance of in this case) that makes it even funnier.
But if you'd like we can just go by 4chan rules here and post: "Fake and gay!"
But the thinking model, you know, also uses a lot more power. The solution is simply to use local models that can think, but of course that has become even more out of reach now thanks to AI companies.
55 Comments
ininewcrow@lemmy.ca · 91 pts · 92d
A river was damned with water and turned millions of gallons of water in order to give us this digital abomination
southsamurai@sh.itjust.works · 13 pts · 92d
Mad love for that word play
decended_being@midwest.social · 43 pts · 92d
https://giphy.com/gifs/huh-steve-brule-dr-kc0kqKNFu7v35gPkwB
IAmNorRealTakeYourMeds@lemmy.world · 23 pts · 92d
I'm rotating Russia incorrectly. you can't stop me.
18107@aussie.zone · 12 pts · 92d
Stand it up on one end?
GreenKnight23@lemmy.world · 7 pts · 92d
Russia goes where?
that's right! in the Japan hole!
anomnom@sh.itjust.works · 3 pts · 91d
Yeah rotate in 3D and it’ll fit nearly infinitely inside its outline. Unless you count the thickness of the crust or something topographical.
pewgar_seemsimandroid@lemmy.blahaj.zone · 42 pts · 92d
whatever ecosia uses
HereIAm@lemmy.world · 27 pts · 92d
To be fair, I suppose it is a well known fact that countries do indeed fit inside itself
Simulation6@sopuli.xyz · 6 pts · 92d
Country usually has coastal waters. For example 12 miles off the coast of the US is still considered part of the US. So the country US does not fit into its physical land area.
sukhmel@programming.dev · 1 pts · 91d
Maybe op asked for physical area to fit into country, since we're specifying ambiguous part
OpenStars@piefed.social · 2 pts · 91d
Not mine though.
pewgar_seemsimandroid@lemmy.blahaj.zone · 36 pts · 92d
grok
Comet79@lemmy.world · 22 pts · 91d
PrimeMinisterKeyes@leminal.space · 26 pts · 91d
Average redditor speak.
meekah@discuss.tchncs.de · 14 pts · 91d
True but at least it just sounds sarcastic rather than astonished that this obvious fact is actually true.
timestatic@feddit.org · 1 pts · 91d
You know, thats where a bunch of the natural language training for these models come from. (But sure Grok is probably even aimed at this type of speech
RamenJunkie@midwest.social · 3 pts · 91d
That population comment has me wondering now. Is there a country that the population of itself, would NOT fit inside.
Like maybe some random tiny European country that produces a lot of people who go live in larger nearby countries, but they are still citizens of their home land. Then for some reason, they all decide to return home.
I guess my point is, that actually could very easily be true, that there is a country that can't actually contain its population.
I mean, physically, probably not, but more like, with houses and shit.
itkovian@lemmy.world · 31 pts · 92d
Truly the technology of the future. /s
OpenStars@piefed.social · 5 pts · 91d
More like the present, unless you want to lose your job, capiche?
itkovian@lemmy.world · 1 pts · 88d
I don't have a job.
OpenStars@piefed.social · 2 pts · 88d
All the more reason to get on the AI bandwagon!
(/s, and I hope you manage to find one, in the case that you want and are looking for one)
NostraDavid@programming.dev · 3 pts · 92d
This isn't the thinking version, which is a LOT better than the instant model. I don't use the instant version any more, due to hallucinations.
fascicle@leminal.space · 31 pts · 92d
Hey at least the profit margin benefited from a few lay offs and market speculation over this
Pothetato@lemmy.world · 24 pts · 92d
You can tell because of the way it is
Mog_fanatic@lemmy.world · 4 pts · 92d
That's pretty neat!
pewgar_seemsimandroid@lemmy.blahaj.zone · 18 pts · 92d
mistral
pewgar_seemsimandroid@lemmy.blahaj.zone · 6 pts · 92d
boonhet@sopuli.xyz · 5 pts · 91d
My Mistral conversation was very similar to yours too so I won't bother posting it
Axolotl_cpp@feddit.it · 1 pts · 91d
Gemini loves to yapp
Edit: i asked it again
SkyezOpen@lemmy.world · 1 pts · 91d
My phone got force updated to gemini and I kind of dig it. Mainly because when I told it to navigate to costco and it asked whether I wanted to go to the bakery or the food court I said "whichever one because they're in the same building you fucking clanker" it just routed me to one of them instead of saying "sorry, I don't understand" like Google assistant usually did.
Realistically though, using an LLM for natural language processing should have happened years ago and I'm pissed it's so enshittified now that it's finally happened.
Axolotl_cpp@feddit.it · 1 pts · 91d
It would be better if the damn LLM was local, i tried to do it on a crappy 300€ phone, it worked, the LLM was one of those stupid ones but it did a deceny job for a 300€ phone, so i can guess that with flagships it would be good, heck, if we had a processing unit dedicated to LLM stuff it wouldn't be that bad probably
Whostosay@sh.itjust.works · 16 pts · 92d
Alright it's been fun boys, pack it up, this shit is over
lauha@lemmy.world · 10 pts · 92d
Still more intelligent that most CEOs
MousePotatoDoesStuff@piefed.social · 14 pts · 92d
China is the most populous country, as well as one of the largest, to achieve this.
This was made possible due to the One China Policy (1 China = 1 China).
timestatic@feddit.org · 10 pts · 91d
Crazy how somebody is even mentioning the fact, that Russia can fit multiple times in itself, but only, and only if rotated correctly. I also love the sentence used to conclude by ChatGPT:
Any Geographers here to explain? But no seriously I know this is frequent but the way it speaks to make its statement seem believable by citing non-existent expert certainly is not one of the better sides of LLMs.
xthexder@l.sw0.com · 8 pts · 91d
No, no, it clearly states Russia can only fit inside itself multiple times if rotated INCORRECTLY. Like, opps, I fell through a 4D wormhole while rotating Russia.
timestatic@feddit.org · 2 pts · 90d
Wait i overread that thats absurf
chunes@lemmy.world · 10 pts · 92d
DeepSeek does a little better with this.
atopi@piefed.blahaj.zone · 5 pts · 91d
they both are nondeterministic
both could give the right answer and then the wrong answer with the same prompt
one try is not enough to say one model is better than the other
GreenKnight23@lemmy.world · 6 pts · 92d
my fuckin sides! has anyone seen my sides??
MonkderVierte@lemmy.zip · 4 pts · 91d
Also Switzerland (if you make it fit).
AI_toothbrush@lemmy.zip · 1 pts · 90d
Duckduckgos ai thingy doesnt fall for it
LuminousLuddite@lemmy.world · 1 pts · 91d
Nomad@infosec.pub · -7 pts · 92d
Faky McFakeface
Its almost sad that people hate on ai for its power usage except when its used to mock AI and mislead the public.
Over 100 upvotes on Lemmy and noone checks this shit? Sheesh you wanna believe, am i right?! ;)
drcobaltjedi@programming.dev · 30 pts · 92d
Okay, your output is different given the same input... So what? It's a well known fact that these LLMs are non deterministic. Theres a guy on youtube that asks chatgpt everyday to count to 200 until it doesn't fuck up. Your output does not prove or disprove the authenticity of the original post.
arcterus@piefed.blahaj.zone · 13 pts · 92d
Tbh them being nondeterministic is a big part of why they're so unreliable. Like, maybe it'll work fine for 9/10 people, but then there will be that one person whose home directory gets wiped for whatever reason. Or maybe it'll do math right for those nine people, but then for that one person it'll say
1 + 1 = 11.You're basically gambling if you don't verify the answers.
FishFace@piefed.social · 5 pts · 92d
Not really... Determinism would only help if you could copy someone else's prompt and history 100%, which you generally would not be able to.
Because maybe it always gets 1+1 correct, but fails 1+2.
arcterus@piefed.blahaj.zone · 2 pts · 91d
I'm referring to nondeterminism for the same prompt, since unless you start a session from scratch, it's unlikely you'll have the same history. If you give it a prompt, then depending on what you've told it previously, it may blow up in your face.
FishFace@piefed.social · 2 pts · 91d
Determinism for the same prompt means you can't give it context through a conversation, which vastly shrinks its utility.
That said, even that form of determinism can be unreliable: the example of arithmetic still works; you could have it completely deterministic, but if it only performs correctly on 80% of arithmetic problems, it's still unreliable.
chunes@lemmy.world · -3 pts · 92d
In fact, if you give it this prompt 50 times and it only fucks up once, that clearly indicates that this is post is misleading. It's also likely this post was faked with a different prompt than the one shown.
Chozo@fedia.io · 15 pts · 92d
Open a new thread and ask again. You'll get a different result. Open a third thread and ask again, you'll get yet another result. That's how LLMs work. You getting a different answer to your prompt doesn't mean anything.
FosterMolasses@leminal.space · 1 pts · 90d
Literally no one checks anything on Lemmy lmao
People will read misinformation on the top comment of a Today I Learned post and take it as gospel truth before they click on the actual link of the post that immediately debunks it.
GeneralEmergency@lemmy.world · 0 pts · 91d
Shhhh. Don't bring logic into a Lemmy thread.
It scares the Lemmites, because it reminds them they're not as smart as they try to act.
Thorry@feddit.org · -1 pts · 92d
We are in a memes community here, most memes are only tangentially related to the truth (if that). It's a joke, it doesn't have to be true, but if it is true (which there might be a chance of in this case) that makes it even funnier.
But if you'd like we can just go by 4chan rules here and post: "Fake and gay!"
NostraDavid@programming.dev · -3 pts · 92d
Just use the thinking model, which severely reduces hallucinations. Only silly people use the instant model to mock a tool for doing what it does.
stephen01king@lemmy.zip · 2 pts · 91d
But the thinking model, you know, also uses a lot more power. The solution is simply to use local models that can think, but of course that has become even more out of reach now thanks to AI companies.