talking to chatgpt is like asking advice to 1000 redditors

>>221868313
People used to think search engines were a gimmick and before that they couldn't believe that a CD could hold as much information as a file cabinet. But don't worry Reddit said it's a bubble and the nazi tech oligarchs are going to lose all their blue chips.

>>221868326
Who do you talk to then? I know it pulls information from reddit, but it does from many sources and sometimes this is useful.

Being able to get the opinion of 1000 redditors in a second is very powerful.

>>221868313
An openai model solved recently a 80-year math peoblem in discrete geometry actually

>>221868313
Claude is better imo, shadjibidi waffles too much
>>221868326
Mainstream redditors are retarded mouth breathers, but 5 year old posts on niche subreddits asking about specific problems are some of the best information sources on the internet if you need a solution

>>221868313
Retarded kiwi AI is obviously useful

>>221868339
I am one of these people and it is a bubble (financially speaking) but the underlying technology is revolutionary. As soon as it gains mainstream public acceptance in everyday life, it makes so many roles/processes obsolete.

>>221868313
Seems like it's flatlining though. Clearly is not getting exponentially better anymore.

>>221868423
AI was winning math olympiads even 2 years ago with meme models like Google Deepmind. That's not surprising at all.

it's very good at giving you wrong answers but written in way that it looks very credible (because you don't know much about the subject, otherwise you wouldn't be asking)
yeah it's useful but you still need to make your own research after asking it, always take everything with a grain of salt

>>221868313
I have AI fatigue

>>221868313
>quality information
Doubt

>>221868499
Weird, third worlders are usually suckers for AI

>>221868313
I think the great filter alien thing is 100% them putting everything into language prediction models and thus spawning retards like OP who think a glorified hallucinating search engine is the future

>>221868477
Okay but there's a difference between regurgitating known things and discovering something new

>>221868521
I'll ask ai how to make myself immortal

>>221868448
I am a retard, but I would still say my tech knowledge is greater than 90% of the general population. 50% of people can't even navigate their phone settings.
>>221868467
From my very limited knowledge, it is already good enough to fundamentally change the structure of society (particularly around white collar/office work). I just don't think the application has really started yet in and there is still too much public resistance. It's going to gut the public sector which is a huge part of every western country (and therefor peoples disposable income).

>>221868513
what was that again I couldn't hear you over the tchiiiiiip sound you were making

>>221868540
>it is already good enough to fundamentally change the structure of society (particularly around white collar/office work
We've had technology to put these job to sleep since the 80s

>>221868490
>>221868516
Let's say you have a question: What are the long term consequences of taking Epilim taking into account my medical history? The only information to be considered is peer-reviewed studies and my own medical history.

Who is going to give a more thorough, cheaper, considered and unbiased answer? A GP or AI?

I understand what you are getting at, but in the right application with the right source data, this technology is mind-blowing. Don't think of it as intelligence, think of it's ability to analyse and report on data. As a reporting/communication tool (which is a huge part of most white collar jobs) this is better than any human and constantly adaptable with new information.

>>221868423
It solved this problem btw
https://en.wikipedia.org/wiki/Erd%C5%91s_distinct_distances_problem

>>221868540
>I am a retard, but I would still say my tech knowledge is greater than 90% of the general population. 50% of people can't even navigate their phone settings.
I think the retarded move now is not using AI to take advantage of fake job havers.
Never underestimate technology again

i have asked it medical questions before and it didn't give me anything useful

>>221868617
This.
>OK AI
>Take these sources
>Stick to them
>Solve this doubt I have
>Provide pages and paragraphs involved in the development of your conclusions
WE HAVE CREATED THE GOLEM, THE BEAST ANCIENTS WISHED FOR AND WE ARE HERE QUESTIONING IT'S UTILITY LIKE A BUNCH OF FAGGOTS

>>221868645
I don't think chatGPT has an easy cure for beheadings, Pablo

>>221868313
Enjoy it while it's still "free".

They're burning a billion a day on this hoping everyone starts relying on it so they can start charging you out the ass for it.

>>221868668
Just run one of the Chinese models locally. As a side benefit they can't steal your data too.

>>221868668
Wait until local open source models are good enough and then get yourself your own rig.
Steal energy o algo.
+Start data hoarding, the plan is to poison the well

File: 1779925746176349.mp4 (3.92 MB, 720x1280)

3.92 MB MP4

>>221868313
.

>>221868724
>Artificial
>Intelligence
Niggas mad clankas became intelligent before they could

60 percent of business has implemented it already. You are a behind if you don't. I find it scary though. Read up on Palantir, they can track you with AI, then drones. I could imagine british police using this in a year or two.

https://youtube.com/shorts/7lXEeqT9LnQ?si=CqCUCsfEDHS8wjJD

>>221868313
>quality information
Not always. Don't let the psychosis get to you like it has all of the Californian and Texan tech elites.
It's going to be a useful tool in the immediate term, and a very integral part of work in the future. That is all.
But the economics of hosting it for people and making money just isn't really there like they hoped. So prepare for mass enshitification.

>>221868821
Or prepare to invest in the local infrastructure for your own like I did.

>>221868622
I agree, the question is how can I learn more to achieve this. Not in a get rich sense, but how can I learn more about this technology.

>>221868361
>it pulls information from reddit
You can tell it not to or tell it to weight reddit responses lower.

>>221868696
How to I go about learning more about this. I made a wow private server once in highschool with friends so I think I can do it.
>>221868668
Naturally, but I would pay for this. I am tempted to dedicate money to it right now if I can find a value case for it, which I believe I can if I can learn more.

>>221868617
>Who is going to give a more thorough, cheaper, considered and unbiased answer?
Doing your own research and filtering empirical information + anecdotes together to piece up a good mental image. Takes time but guarantees intrinsic knowledge
Nothing mind-blowing about skipping all the above to give you a milquetoast summary that depends on your digital footprint to please you and stroke your ego

>>221868313
Most of these answers are 80% lies. I'm not even AI hater, I dont generally care, but every time I asked it for help with such and such it just lies.

>>221868791
Old farts are easy to scam, you will be "behind" temporarily on menial task automatization but on the long run humanchads win

>>221868313
>be 30 something chud
>believe AI is an overhyped gimmick
>start using AI for porn, both 2D and 3D
>Now im a porn addict
Scary tech

It's not even an AI, it's a predictive token generator that guess the next character. That's literally it. It doesn't think, it doesn't understand anything. AI is genuinely going to kill the entire internet and the tech sector entirely before this grift ends. It's not even a viable business model the AI companies lose money because it costs more to run than what they make - right now they've admitted they're only charging about 3-5% of what they should per token. Massive investment scam.

>>221869011
It has not clicked for you yet, it is why you don't get it. AI is doing exactly what you are describing yourself manually doing, only in a much more efficient and faster manner. You can still review the sources or prompt more. You are making the mistake of assuming you (as a human) are more intelligent and unbiased at data analytics than this technology, which is completely incorrect. Humans are inherently biased and emotionally driven. Your own response is proof of this.

Assuming "doing your own research" means consulting university publications/medical journals directly and studying the issue intrinsically, alongside speaking to a wide range of people with anecdotal experience (and not just googling the question and reading a few reddit posts), AI will be able to do this in seconds with the source material. All those medical journals and anecdotal experiences are just data points you can get the AI to use.

When someone has the ability to upload all medical journals and medical notes into a isolated AI solution and sell it back to medical institutions as a diagnostic tool, they will make billions. This will (and probably already has) going to happen. INFACT, you just reminded me that last time I went to the GP, the doctor literally used Gemini AI in front of me to get guidance on their medical advice.

Lots of people (particularly in high "intellectual" positions) stand to have their whole lively hoods challenged by this tech, so there will be resistance and fear-mongering, but at a fundamental level it is already apparent this is going to wipe out most "knowledge" based roles.

>>221869269
Yes financially there is a huge grift behind this phenomenon, but the underlying technology in exactly the way you explained is revolutionary.

I can see in this thread there is a trend with some to literally view this as "Artificial Intelligence". Yes, it is 100% not this. The ways in which data can be analysed more efficiently, accurately, without human biased and then reported back via a simple GUI in adaptable terms is ground breaking though, particularly in industry/government environments.

>>221868313
same. i used to think that IA was just a overhyped "deluxe google" but since i got a new job i've started using it several times a day

i use AI for academia - the humanities, I pay for the best Anthropic model, it allows me to work faster, yes, but even then it keeps hallucinating shit and making up citations and things that simply do not exist.

if it fucks this badly with just text i can't imagine what it does with code or things that actually matter like calculations for, say, bridges or a building's structural integrity. scary shit

>>221868313
I've noticed multiple times that it gets facts wrong, people like you are the reason the average IQ is dropping

File: get a load of this retard.jpg (243 KB, 720x750)

243 KB JPG

>>221868313
>quality information

>>221868313
the only reason you think AI is great is because you're genuinely a midwit

>>221869734
>>221869805
>>221869809
its getting better doe

>>221868326
You can tell it not to.
>>221868437
ChatGPT is good as a general one. Claude is better at psychological or philosophical or generally nuanced queries and questions.

File: gaza-facial-recognition-uk.png (390 KB, 1000x964)

390 KB PNG

>>221868791
it's already being used

>>221868313
>Chatgpt
>2026
Unc you're late. Use gemini or claude, they blow chatgpt out of the water.
>>221869321
The thing is AI is a massive bubble, no AI company has been able to turn an actual profit yet. AI will have it's niche applications in some fields, but by and large, it will lose it's ubiquity.

File: HIfMNU2WsAAs7CW.jpg (46 KB, 500x574)

46 KB JPG

>>221868313
i had a problem with it hallucinating and thinking that some redditior's ironic comment was a legitimate source of information. if i didn't know any better, i would have accepted reddit comment vomit as fact too

I have written 36 novels using AI.
I have an API which I use to make Claude write passages, then have GPT 5.x review them and then Claude makes changes. Like a writer and an editor sort of deal. This is all automated.
I give ideas and revise the whole thingie once it is finished.
Already made back what I spent on models through Amazon book sales.

>>221868313
It's a tool of fascism and we need to stop it before it takes the jobs of well-meaning left wing people.

It's already making people retarded. I weep for the future generations.

>>221868339
>>221868460
Yeah it's like the dotcom bubble in early 00s.
It'd get accepted eventually in a largely diminished but streamlined state after the hype died down.

Opus is the only model that impressed it, I've subscription for both, Opus clearly peak higher, especially for very long context tasks, but more inconsistent overall.

>>221871625
>I have written 36 novels using AI.
AI has written 36 novels for you

They're fun for a little while to fool around with and brainstorm, but even then the hallucinations start happening. Can't imagine using it for more serious work but apparently non LLM AI's have been able to detect illnesses in the medical field better than humans.

>>221871625
I use them for free coverage of my work instead of paying for failed novelists/screenwriters to give me generic review for $300. they legitimately can't write shit independently unless you guide them into it, which at that point you already did all the work.

AI is only good at deconstructing and feedback, Re-affirming the concerns you probably already had. utterly irrelevant when it comes to creating stuff.

just asked Chatgpt and it agrees with me it's useless a lot of the time and mostly relied on by retards

>>221872051
Which is why i'm actually bullish on the effects of AI, especially on the film industry, the sheer volume is going to eventually force selection toward originality and voice rather than generic competence.

File: 1776394059335997.png (60 KB, 250x250)

60 KB PNG

I just asked it if there's an office that does same day Invisalign retainers and it just completely lied, kys

>muh hallucinations
How do you fail to operate AI in 2026?

>>221868313
Like what?

I think I used up my allotment of Perchance because it doesn't seem to be working today. It seems to be the only one that doesn't require an account.

>>221873435
What one? I like trying different types out.

>>221868326
That's a good enough working model of how to use it yea

>>221868313
I would still use Google if it weren't fucking useless these days. Before, you'd search for something and the first page would mostly be forums or blogs with worthwhile information; now, the first thing you see are sites that have paid to be at the top (ignoring Google's AI), and even going to the second or third page isn't helpful.

>>221868313
the possible impact of AI+robotics is not overhyped.
Its widespread use depends on backend readiness which includes data centers, energy sources, application adaptations ... It takes time to prepare all this.
I have no idea what they are going to do with humans. Especially in cunts with high protest potential. setup fights migrants vs locals? or start another ww? endless options are available...

>>221868696
Local models are garbage for anything useful and the barrier to entry is still an RTX 5090 at most (I have one, not even the latest dense Qwen model is decent).
>>221869734
>I've noticed multiple times that it gets facts wrong
Damn, just like human beings. Crazy!

Paradoxically the value of AI is entirely dependent on the expertise of the user. So far about 99% of of people use, interpret and value AI completely wrong. Thusly I am confident in saying that AI will statistically be a detriment to all who use it. In the end it's just a tool, a brush in the hand of Trog or Picasso.

Ai is good until you ask it something you are knowledgeable about. Then you see it shit out nonsence

>>221868313
careful
it can hallucinate
double check everything

>>221868313
I stopped trusting the answers when I realized it makes a lot of shit up just to please you. Better not get used to it.

>>221874155
Yea it feels like so much of IT has been an arms race. Probably why people shit on GDP these days

>>221868313
Got a NEET friend obsessed it will take the jobs out of the rest of us who work in IT.

I told him several times that a gret part of coding involves maintenance of existing software to meet new requirements, and AI-made code takes way longer to work on.
He spent like a year building a graphic adventure using claude. AFAIK, it was fully coded in JS and the assets gobbles up like 6GBs. I haven't tried it yet, but some guys have told me it locks up at the begining, and it's plagged with glitches.
Good luck finding&fixing the issue.....

>>221868326
Personality of a redditor definitely with modern values

>>221876002
Oh there will be issues. Catastrophic issues, even. It's still going to take our jobs and do an infinitely worse job for a 0$ salary.

>>221869321
>only in a much more efficient and faster manner.
>When someone has the ability to upload all medical journals and medical notes into a isolated AI solution and sell it back to medical institutions as a diagnostic tool, they will make billions
And will get millions also killed because most scuentifical papers are fraudulent

>>221868313
you have to be really specific with your prompts if you want to do more complex tasks and I don't like writing long prompts. maybe in a few more years these AI systems will work with lazy prompts to one shot the exact output I want. But they are still incredibly useful tools and will 10x the economy once people figure out how to leverage them for their jobs

>>221877246
that being said I will never use them for life advice and other personal stuff because AI is still sycophantic and will tell you what you want to hear

>>221877275
What's the alternative?

>Tell me the nutritional contents of a banana
>ChatGPT: No. Fuck you.

>>221877343
subjective advice. not objective facts like nutritional content. they're good at that stuff

>>221876998
It's not free. Tokens are expensive. And will end up getting 20-25x more expensive once angel investment money dries up. Be lucky if any models outside of the Chinese open source ones even exist once the bubble pops.

You can make things with it that it would take hours. Even if you have to review the research that it does because it does hallucinate a lot. He fact that you can use it to do that and list their sources is really good. People who don't use AI are being stupid af IMHO.

Welcome to the future. As a code monkey I feel like I've been given 3-5 slaves to code under me.

>>221872051
This is right. You have to push it towards where you want to go and you also have to feed it things. We are still far from "no more programmers in 6 months" and those things are good at it.

>>221878552
Yes. I'm thinking about not even looking for a job and just doing my own thing.

Everyone moved to Claude 6 months ago

>>221869269
>It's not even an AI
That's not really true. The term AI has been applied to lots of different technologies for a long time, and doesn't literally mean it needs to be intelligent like a human. Even chess engines that use maths to predict chess moves are considered AI, as are computer vision models that are also just maths.
I also don't know where you got that figure about them charging 3-5% of what tokens are worth. Most of the frontier labs aren't profitable yet, but the cost per token (for them) has decreased by almost 1000x in the last few years, so they're probably not running on very bad margins by this point.

File: monke user.png (653 KB, 808x610)

653 KB PNG

>>221868313
It's better than search engines these days for sure. Google is borderline unusable, and its own AI embedded at the top is garbage. Still, I disagree that the presented info is of any good quality. I double-check everything because there's so much hallucinated garbage it's not even funny. But that might be because Polish legal acts and their interpretations are largely paywalled and embedded in 3 pages deep downloadable untranscribed PDFs.