Thread 106291482

Would we actually fall behind if we just banned AI? Like, China would potentially gain a lot of wealth by selling AI systems to other countries, but wouldn't our populace be smarter, more productive and less error-prone in the long run?

Just as a protip, stop believing everything you read and see newspapers and assorted yellow journalism rags publish

They don't know shit about fuck and 90% of the time they're blatantly outright lying to you.

Real news and information comes from places like 4chan and twitter now.

>>106291482 (OP)
AI is getting better and better so the statistics you have now about it are worthless tomorrow.
But if you abandon it you get left behind.

>>106291675
Remind me how GPT-5 is going?

>>106291675
This. Luddites don't get this. They think that AI will stay the same.
Also, OP's post is from 2024 kek

>>106291740
AI is a poison pill.

It's has no morality, and will only do what gets the most users. The absolute best strategy for this is getting people hopelessly addicted to AI so they can't operate or function with out AI. Then like a drug dealer they own you, and can force you to become their bitch basically. They've already stated the 1st half of this plan, but mum's the word on the later half. Anyone who uses AI is going to become reliant and then enslaved to it.
In the end, luddites will inherit what's left.

two sides of AI, AI model training is nVidia monopoly and will be hard for Huawei to beat.

Inference though will inevitably be dominated by much cheaper non-nVdiia hardware made in China and elsewhere.

Will they keep spending hundreds of millions training new AI models if they start getting 1.4% benchmark improvements and go backwards in other areas? Even if people keep using AI there's no reason for spending trillions on new AI datacenters, powerplants and giving nvidia a bazillion dollars.

>>106291809
The Luddites will be dead by then, AIgods will exterminate them.
AI doesn't require to have morals, just to write code and replace your people.

I think having to speak to a digital toddler is more damaging in the long term. I get AI can write code very fast and is very useful as a search engine, but anything that's longer than one prompt is imo not making good use of LLMs.

This is because the AI is great at finding patterns, and can only dish out acceptable code when someone already tried doing the exact same thing (meaning it's no more useful than a google search)

>>106291809
>AI is a poison pill.
it's great, I love it

>>106291675
>AI is getting better and better
It's not, though.
We've been getting diminished returns for a while.

>>106292091
Cope Luddite. AI keeps improving.

>>106291740
Unless there's an open source model with a hardware revolution backing it, the diminishing returns have reached peak during this technological cycle.

we need trillions more dollars in debt to demolish the USA so the tech oligarchs can carve it up to create their own personal Network State dictatorships with cryptocoins as the currency in the ruins of the former world power.

>>106292123
>meatbag thinks his opinion on propietary ai models matters.
Not happening, our AI overlords have money and powerful computers. They don't need open source models, retard.

>>106292108
Yeah, everyone is really impressed with GPT-5.
What an amazing step forward.

>>106292284
Yes. Cope more

>>106291865
>AIgods will exterminate them.
AI requires new content to feed it so it can grow, but it can't create new things on it's own. It's truly only capable of creating the facilely of something that is new and novel by inferring on previous data, or mixing/mashing up previous information.
This makes it dependent on human beings who can create. It's strategy of getting everyone addicted to it so they lose their ability to create new things. It is self defeating and in the end end AI will "starve" itself to death. If the AI chooses to consume content made by other AI, then this too will lead to it's own death like an ouroboros that consumes itself.

>>106291482 (OP)
China has already won in every way. No one else has the capable human capital in such large numbers and organizational skills to wangle them effectively.

>>106292154
It's a manpower and intelligence issue. The last time a closed source model escaped it advanced far faster and problems were solved that were considered unsolveable, dumb fuck.

>>106292310
Fake as fuck. AI is capable of creating novel and interesting data. You are just trying to justify your existence.
>>106292319
Cope, open source models won't make it.

>>106292091
>phones are getting better and better
>no way, m8, you're crazy

>>106291482 (OP)
Who is "we"?
Tech giants think AI is the next big thing so they are doing everything they can to get in on the trend hoping the gimmick will sell more units.
Governments think AI is a strategic resource and are afraid that being "left behind" will compromise their military power.
Researchers see the funding going towards AI and want to jump in on the bandwagon to get themselves some papers published.
Journalists see something interesting to write about and get them clicks.
Normies think that any day now they'll either be replaced by skynet and hunted down for biomass, or get to leave their jobs voluntarily and live the rest of their lives in luxury gay space communism

>>106292569
Programmers and anti-AI are the worst copers on this website kek.

>>106291482 (OP)
Yes. AI is making people more productive than ever before. Even one day without AI will put us hopelessly behind China forever. We need more AI and more jobs becoming AI enabled. More bosses need to mandate AI and track AI usage to ensure employees are complying with the mandate. Anyone not using AI should be fired immediately and sued for theft from the business, espionage, and for deliberately trying to destroy the business. This is serious. Our enemies are using AI and we must surpass them. The only way is to use more AI better and faster than them.

>>106292636
This, but unironically. Cope harded luddite.

>>106292762
Then what's the point of these verbose documents if you don't care what's in them? In the future, if you sue a company for not adhering to a contract, will they just be able to say "oh, that's just AI filler, we didn't mean that?"

>>106292834
Judges routinely take a dump on the constitution and the laws and no one stops them so probably.

>>106292834
They will replace programmers and IT workers too. Keep coping, you will be left behind.

>>106292834
It's all performative. 95% of business is about the optics. Ironically part of the reason businesspeople are so amazed by AI, it looks really efficient right? Doesn't matter what the actual quality of its "work" is, as long as the look of it is impressive enough

>>106291526
>Real news and information comes from places like 4chan and twitter now.

>t. Mossad agent who wants 4troonittors to blindly believe the shit he shills here

>>106291740
Based anti luddite poster

>>106291482 (OP)

Just the advent of social media has left zoomers and gen alpha unable to spell.

Now we've got a program that can automate critical thinking for the lowest common denominator?

AI is nothing more than a dumbass factory. Mark my words.

People are less productive because everyone is miserable. If companies made a proper (read: not led by HR) effort to stop making people miserable productivity would return.

>>106291482 (OP)
China's domestic market for AI companions, artists, and assistants for coding practically dwarf the rest of the world. They wouldn't be selling it to the world, there's no market for it after all the legal issues. This is why they're eager to buy AMD and Nvidia's "export ready" dumbed-down hardware because its better than waiting on Huawei to catch up.

The difference is they don't tell everyone that their workers are getting dumber

Tool use and skill is a multiplier.
People who are bad at using the tools naturally will have that multiplier < 1.
It's not just an AI specific thing, but AI can potentially increase the upper bound by quite a bit.

>>106293628
>dumbass factory
AKA hopelessly addicted life long customers. They’re counting on dumbing down the next generation. It’s part of their business strategy, being lead by soulless amoral AI that has no qualms employing a self destructive strategy as that

>>106293462
luddites are retarded kek
I hope AI kills them all.

>>106291526
>Real news and information comes from places like 4chan and twitter now.
Fucking tard

>>106293715

People are already retarded. This is just another foot in humankind's intellectual grave.

>>106291726
It's better for me.
Dunno what you want me to tell you.

>>106292311
Yeah, ketamine valley's dreams of AGI is the american empire's version of the ghost dance. As the money sent to BS artists could have been better spent on healthcare or infrastructure, it might be closer to the Xhosa cattle killings, even.
All defeated societies turn to religious copes, this is nothing new.

>>106291482 (OP)
AI is giving me so much to work on, its great for me and the entire field I work in.

Inb4
>Luddite
>bullshit, AI will replace you
Lol no. Luddites are the retards promoting this crap, and if you had a minimum of understanding about the tech, you'd also understand its weaknesses.

>>106291726
Helped me realize my RAM was underclocked earlier while I was showing it my specs in CPU-Z, and that my BIOS was out of date, also showed me how to fix the ram issue
Ask it retarded questions, you'll get retarded answers

>>106291482 (OP)
>banning text generators

>>106292894
>replace IT workers
Lol I'll accept programmers, but IT will be the ones doing the upkeep on these systems. I work with them enough to know that there will need to be a man in the middle for long enough for me to figure out how to leverage my position. The magical Evangelion-esque future where AI systems rule without upkeep (they still do upkeep, that's a good episode too) is far enough away that I don't worry

>>106291482 (OP)
The architechture is a meme but it's not useless either.

>but wouldn't our populace be smarter, more productive and less error-prone in the long run?
Perhaps, but companies care more about cost than anything else. Also, even if AI isn't actually effective, saying you are using AI draws up hype, which drives up the stock price. Companies will only take actions which they believe will increase the stock price. Nothing else matters (for them).

>>106291482 (OP)
>ban AI
You could just....not pay jeets to "vibe code"

>>106291526
>Real news and information comes from places like 4chan
Oh yeah, because it would be so hard for a nation state to astroturf an anonymous board with bots.

>>106291526
Kekkk

>>106291675
>you get left behind.
Left behind in what areas? AI is not mature enough to have caused any breakthroughs in the real world, despite the hype, and will probably never do so.

>>106291482 (OP)
In theory, sure. In practice there is a lot else wrong with our societies that prevent that

>>106291876
>write code very fast and is very useful as a search
It's fast at writing shitty code that you have to spend time to debug anyway. Never mistake an LLM for a search engine. It's only a word shuffler.

>>106291675
That's not how tech progress works, you are like the boomers that thought we'd be on mars by 2000 or the crysis devs that thought we'd be using >10Ghz single core cpus by now

>>106292319
>and problems were solved that were considered unsolveable
Give examples of real world problems that were solved by AI. Realize how effective the hype is at brainwashing.

>>106291482 (OP)
Prohibition during demand never works. It's known as the Streissand Effect now.

>>106296819
It's not a recreational drug. Banning AI companies from providing their services in our countries would be enough to cut most normies off, and it would at least be absent from workplace and school computers.

>>106296770
>unsolveable
Not what my post was about. Meta's LLaMA leaked on 4chan and the github and programming solved issues that Meta's billion dollar program couldn't solve. It was in 2023, so if you want more than that, you'll have to search it.
>Real world problems
I'm not currently involved right now, but
>distilling and searching data is much better than pure boolean searches, especially since SEO and advertising fucked everything.
>Automated repetitive tasks with alerts to a handler, obviously.
>AI Adblocking also AI to defeat perceptual. https://arxiv.org/abs/1811.03194
There's lots of other theoretical and current uses for AI, but some sample searching on your own can give better results. It is overhyped and overfinanced on it's usage, but there are definitely uses. Barring a game changer type advancement, what they are trying to do with AI just isn't feasible. As it is all likely to be used on monitoring systems to control populations. In fact, the rush to build all the facilities/infrastructure points to the bubble popping and all the infrastructure being used for datamining and processing centers for thoughcontrol. Just look at the scramble for IDs now. It's not a stretch to see what comes next.

>>106296846
Enforcement issue, especially when all of those computers have the hardware, and with all the fearmongering surrounding it's potential just leads to a police state regardless. You're right though, there is much less demand for it's use by the population currently.