[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/pol/ - Politically Incorrect


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: images-23.jpg (21 KB, 474x421)
21 KB
21 KB JPG
>4 KB of memory
>1 MHz processor
>floppy disks
This was state of the art less than 50 years ago and I'm not supposed to be afraid of AI in the future?
>>
if you are the "creative" field, you should consider jumping in front of a train
yes that goes for coding as well, its a creative field
>>
>>517687561
Literally nothing changed in 50 years, the transistors just got smaller. No massive leaps in any tech. Not even monitor technology- again, the lights just got smaller. It's still tiny switches and tiny light bulbs. Cars still run on explosions. Spacecraft still run on explosions. Medical knowledge didn't even really advance, they just were able to centralize access to the info.
>>
AI is great but we're gonna lose a lot of what makes being human special because of it. I don't care because I'm past my prime.
I'm uneducated and I was able to use the free chatgpt to create 2 functioning computer programs. One of them is used every day by me and my coworkers.
there will be no comfortable, decent-wage jobs for average guys who studied and learned python or whatever
>>
File: ChromePolish.webm (2.5 MB, 1920x1920)
2.5 MB
2.5 MB WEBM
>>517687561
Hard to predict, but the world really does seem to be passing a moment where you need to either get rich now, or you'll stay poor forever. AI in the workplace is the ultimate ladder pull.
>>
File: x.gif (1.79 MB, 640x504)
1.79 MB
1.79 MB GIF
>>517687561
>I'm not supposed to be afraid of AI in the future?
You're supposed to be afraid of it now.
>>
>>517687561
Software were simpler back then, but you basically could do everything with 64k of ram that you can now do with 8gb of ram. For the few tasks you couldn’t do there were more expensive machines capable of doing that.
>>517688755
Exactly. I could write a letter on an apple II, use a spreadsheet, code in different languages, write on a modem chat, play exciting games. If internet was there back then I’m sure I’d be able to navigate those websites, with a much lower resolution and black and white but still pretty much functional

Overall, it’s incredible how much modern software provides to achieve only very marginal improvements
>>
>>517689172
for creative bullshit jobs (including coding)
There is still going to be demand for brick layers, welders, health care, different services, agriculture, security... just not creative jobs and no demand by those that held them, because they are going to kill themselves
>>
>>517689833
How are people going to pay for homes and health when the creative jobs go away? Everything is connected, and there will never be enough infrastructure workers to just pay each other.
>>
>>517690004
structural deflation is a given looking at zoomers and millenial birth rate. one or another way. nobody is going to give a crap about art fags and coders ending under bridges or suiciding. they re not needed anymore
>>
>>517689401
i'm afraid of it because it supports my belief that reality itself is some ai-generated truman show bullshit.
>>
>>517687561
My argument is that AI systems face a fundamental ceiling that prevents them from surpassing the best humans. The basis is statistical: human ability follows a bell curve, so most training data reflects average performance, with only a small fraction coming from exceptional individuals.

From this, I argue that AI cannot transcend the upper bound of its training data. While it can aggregate patterns across millions of examples and operate with greater speed and memory, this is efficiency, not superiority. The key distinction is between optimization; applying computation to well-defined tasks like chess or protein folding; and authentic intelligence, which requires understanding, insight, and creativity.

At best, AI can approximate the ceiling represented by top human practitioners. When it explains quantum mechanics or writes code, it is not outdoing the best physicists or programmers, but pattern-matching from their work. AI is therefore “the best average human with superior computation”: an aggregator that combines the strongest patterns in its training.

The deepest limit is agency. Breakthrough thinking requires the ability to resist patterns and pursue unlikely paths out of curiosity or conviction. Without autonomous motivation and judgment, AI remains reactive, confined to optimizing within its statistical landscape. Because agency is not just absent but architecturally precluded, this limitation is permanent.

>t actual AI and Robotics fag
>btw I have a PhD in Math and AI
>>
>>517690602
>I have a PhD in Math and AI
>appeal to 19th century authority fallacy
lol
the books and papers you consumed are in the training data, you are for the most part worthless. also this reads like llm output
>>
>>517690812
>the books and papers you consumed are in the training data
So are the papers and maths I've created.

Prove AI is not largely at best an average human with superhuman computational/memory abilities.
>Fun fact you cannot.
>>
>>517687561
I had an Apple IIe, I spent a lot of time playing Ultima 2 and Time Zone while my older brother played Wizardry and Elite. He even geeked out and was napping all the dungeons on graph paper with his gf who would play with him as well. Better times.
>>
>>517689633
you can't view 4k movies of French girls sucking off Nafris on an Apple II though
>>
>>517687561
My boomer parents didn't buy us a computer so I was miles behind everybody until pirating became a thing and I was on that early because free.
>>
>>517690912
>AI is largely at best an average human with superhuman computational/memory abilities.
In fact this explains "ai hallucination" cleanly because if AI is an average human intelligence, than by virtue it will produce incomprehensible results because the average human intelligence doesn't actually understand anything and only regurgitates matched patterns.
>>
>>517691048
I miss the old days man.
The world changed way too fast. It's nothing like my childhood. If guys from the 50s felt alienated by the 90s, they had no idea how distant the 80s would feel to the 2020s.
>>
>>517688935
>AI is great
It's slop.
>>
>>517687561
>floppy discs

Nigger, you posted a system that is using cassette, floppy discs were lightyears ahead of cassette in terms of functionality.
>>
I still have mine.
>>
>>517690912
>maths
go back
>>
Moore's Law is now invalid, there is a limit to how far we can go and current developments are just trying to scrape that extra 0.001% out of something.
>>
>>517690912
its not a human larp bot, its if you mean various llm models, likely transformer based, in colloquial terms an autocomplete that aligns output to input context based on training data and weights. What it did, and why its making you obsolete, it effectively proved nomnalism and falsified platonic realism
>Math I created
so you patchworked a few axioms together and called it truth, lol
>>
>>517691259
>Oh no he used maths
It's maths not math. Cry about it faggot.
>>
>>517689633
Dude you’re a retard that doesn’t know anything about computers
Stupid faggot frog
>>
>>517689633
>Overall, it’s incredible how much modern software provides to achieve only very marginal improvements
it's because they don't give a shit about user experience anymore all the improvements went into monetizing your engagement, maximizing the metrics collected, and playing pysch tricks on you to keep you glued to it. Modern software is downright hostile to the user
>>
>>517691177
There is no hallucination. that is your bias, the next token being output is still statistuically correct, it might just not align with your cognitive should understanding of reality
>>
>>517691048
I had at them at school and played Oregon Trail on them. The original Macintosh was already 15 years old by that time. I got into C64 programming which uses the same CPU and wrote a 6502 emulator in C.
>>
File: 20250930_154930.jpg (141 KB, 1007x993)
141 KB
141 KB JPG
>Highly speculative & unable to prove it
>>
>>517691489
positivism is religion. Popper had one valuable idea and that was falsification
>>
>>517688755
>transistors just got smaller

The transistors are SO SMALL now that you cannot even see them with an electron microscope anymore. This is a giant leap.
The bigger leap was making a microchip at all though. Look at how sophisticated the intel 4004 actually was. It already had transistors so small you couldn't see them with an optical microscope. And compared to what we have now it was like sticks and stones. A microchip by default is already amazing, and the 4004 was only made to run a calculator.
>>
>>517691617
>The transistors are SO SMALL now that you cannot even see them with an electron microscope anymore.
proof it's a work of the devil and/or other malignant ethereal entities
>>
https://m.youtube.com/watch?v=MnR8DIR5Rx4&pp
>>
>>517688935
>there will be no comfortable, decent-wage jobs for average guys
Political implications of this?
>the tap signing meme becomes more real every day
>>517689833
>There is still going to be demand for brick layers, welders, health care, different services, agriculture, security.
I wouldn't mind to become a shepherd or a beekeeper or a cheese maker or something that will take me out of the city and into the country side getting enough money for a non goy slop plate of food. But how can I become a shepherd at almost 50?
>>
>>517691294
>why its making you obsolete
Lol, wat? It's not making anyone obsolete. If it did, AI would have already figured out agency (choosing to stray from patterns) But it doesn't and cannot because you can't actually program a chip to do anything by it's own choice. Chips follow deterministic or stochastic patterns without self-chosen goals. Nigger.

Second fun fact: My position is widely accepted as fact in AI circles.

>so you patchworked a few axioms together and called it truth, lol
My position isn’t patchwork reasoning; it’s built on a statistical model of training data and provable limitations of pattern-matching systems. In the academic context, these need to be expressed with rigor (formal models, proofs, or empirical validation), not just deduction from assumptions; which is why I can distinguish my view from casual axioms.

>>517691442
>There is no hallucination. that is your bias, the next token being output is still statistuically correct, it might just not align with your cognitive should understanding of reality
and that’s precisely the point. because it is only statistical prediction, it never truly 'breaks' from its training distribution. What you call 'not aligning with reality' is exactly what I call 'hallucination'; a byproduct of pattern matching without agency. The system isn't choosing to be wrong or right; it's following statistical constraints.
>>
File: Indian Wiring.jpg (116 KB, 634x422)
116 KB
116 KB JPG
>>517689633
>Overall, it’s incredible how much modern software provides to achieve only very marginal improvements

Jeets make spaghetti code, it's layers and layers of fixes to fix fixes to fix fixes, layers of unoptimized complexity. For some reason we've decided to allow jeets to write code in our software.
Windows 95 did 95% of the important stuff that windows 10/11 can do today. It could execute and run on a 16mhz 386sx and 4mb of ram. That's how optimized it was. I've run it in 86box on that hardware and although slow, you'd be very surprised that it wasn't slow enough to be unusable.

Pic related is jeet electrical grid. It's in their nature. This is modern software, this is how it's being made now. It's like pic related.
>>
>>517691976
it doesnt require agency, it requires input. A cats movement is sufficient input for it to exist in the ontic sense.
link phd thesis, it would be fun to try to falsify with an llm

>a byproduct
The transformer remains a blackbox and the pattern matching is believing paeno was right. At the current stage the edge cases are not explainable with classic statistic models
>>
File: Subway_Fight_End.png (498 KB, 853x480)
498 KB
498 KB PNG
>>517687688
What like neo?
>>
>>517692531
>just dox yourself on 4chan bro
No thanks.

Your arguments aren't worth addressing because you lack fundamental understanding of what I'm actually saying. You should reread carefully. But I will do it anyway.

Transformers aren't black boxes; they're deterministic linear algebra, and their 'edge cases' are just high dimensional statistics at work. And unlike a cat, which moves from internal agency, an LLM is purely reactive: it maps input to output without self directed existence.

>inb4 animals have no agency
kek
>>
>>517692836
>>just dox yourself on 4chan bro
>>517692531
besides, you can find my papers simply by searching my argument online.
I'm the only one who's actually written on this topic.
>>
>>517691048
>Wizardry
fuck yeah, written in Apple Pascal. Hard to believe it would spawn the entire JRPG genre and all their tropes when its sequels proliferated in Japan.

Wizardry, Castle Wolfenstein, Aztec, Bolo, and Sargon ][ were what I actually did as a school "computer assistant".
>>
>>517692836
The cat movement as input is sufficent for an llm to exist in the ontic sense. it doesnt require humans beyond building "it"
And if you consider "straying" from the array of possible prediction of a model as agency, thats a pretty low bar for agency
>>517692914
KEK pride is a sin. thanks for the dox
>>
File: 1740413630093706.jpg (1.98 MB, 3072x4080)
1.98 MB
1.98 MB JPG
muh AI



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.