Thread 107167233

>a machine will never do arithmetic faster than a human
>a machine will never translate between languages
>a machine will never be able to solve algebraic equations
>AI will never recognize speech
>AI will never play a full game of chess
>AI will never be able to see or recognize objects.
>AI will never be able to play chess competently
>AI will never be able to understand stories or context
>AI will never be creative
>AI will never be able to beat a grandmaster at chess
>AI will never be able to beat the world’s best chess player
>AI will never be able to beat a grandmaster at Go
>AI will never be able to beat the world’s best Go player
>AI will never be able to hold a natural conversation
>AI will never recognize faces accurately
>AI will never be able to pass a high school aptitude test
>AI will never be able to pass the Turing test
>AI will never drive a car safely
>AI will never be able to pass a college aptitude test
>AI will never be able to win international math competitions
>AI will never be able to write publishable academic papers
>AI will never be able to do research as well as a first year PhD student
you are here
>AI will never be able to do research capable of a PhD graduate
>AI will never be able to autonomously discover and test new materials or drugs.
>AI will never be able to produce Nobel worthy research
>AI will never be able to invent new scientific paradigms
>AI will never be able to master common human sense
>AI will never be able to create original art that moves people
>AI will never be able to direct an award winning film
>AI will never be able to run a company successfully
>AI will never achieve super human intelligence
>AI will never be able to govern responsibly
>AI will never be able to achieve consciousness
>AI will never be able to surpass human intelligence in every domain
>AI will never be able to redefine what it means to be human

>>107167233 (OP)
>AI will never be creative
>AI will never be able to hold a natural conversation
These are still true though

>>107167233 (OP)
silence, street shitter

Because hopium overdosing is human nature
>well we may be in a communist UBI Palantir AI dystopia but at least I was right!
t. you by the end of this decade

all we have to do is add a "op will never get laid" right after the "you are here" and we will prevent the dystopian shithole future

AI WILL CONDEMN YOU TO SLAVERY!! IT'S THE WEAPON OF NEW WORLD ORDER!

>>107167233 (OP)
>>AI will never drive a car safely
Still true.

>>107167244
>>107167633
thats the point retard. by the time they're "true enough" to count, you're already fired and your opinion is worth less than the electricity running the model. the "never" list is just speedrun categories for whoever wants your job.

>>107167233 (OP)
when will AI make Yuru Yuri real?

>>107167233 (OP)
It can’t build a house, so it can’t do anything.

>>107167233 (OP)
Silence, machine.

>>107167233 (OP)
>>AI will never be creative
there's no objective definition of creativity thoughbeit

>>107167633
>blah blah blah don't bring up the millions of miles driven of AI vs human
it's ok to be wrong

>>107167233 (OP)
Didn't read.

>>107167233 (OP)
At least half of those are wrong nigger. Also it’s interesting how the people who said that by this year we would have AI agents doing most white collar labour were also wrong.

>>107170215
I actually read them and it’s funny how almost all of the things you listed we already passed were just stuff primitive computers did or playing board games. AI still can’t safely drive cars and still can’t do research as good as first year phd student. It can’t even do proper master thesis on most subjects, I know because I tried it 3 months ago when I was starting to work on my own.

>>107167559
>Communist dystopia as the bad scenario
Lol, that’s the good scenario, the most likely scenario is that once we get AGI for jobs and good robots (we are closer to those then AGI btw) all the non-billionaires just get culled off, and once there is no law enforcing majority holding the survivors back they will fight eachother to death in anarchy. It’s not because they are all evil, it’s just that if one of them if he will try to get rid of the rest so he can become the true and only god of Earth owning all of its resources.

>>107170230
In 100 years which is less than the lifetime of some humans we went from a primitive triode made from melted glass with a few bits of metal rudimentarily put together to H200s, and you're not impressed? Yes we went from electrical transformers being cutting edge tech to transistor gates just a few atoms thick, but now all progress is suddenly going to halt. Sorry guys, that's it lole! Enjoy your 5 fingers! Oh wait, that already isn't a thing. Fuck. Well ChatGPT didn't write my whole master thesis so that means it's over!!!

>>107170328
LLMs stalled 2 years ago, they have not improved beyond meme benchmarks.
>Muh smaller transistors
Yes there are periods of development in a field and then stalling. Each technology development follows the S curve until we reach another breakthrough, but breakthroughs are random so you cannot count on them. Also since you mentioned the few atom thick transformers, how do you go smaller then that? Again, classical hardwere will eventually hit the end of the S curve. The fingers meme stopped being a meme the same year it started and image generation reached hyper-realistic levels in 2023 and then once again stalled. Stalling of AI development in different sub fields is not just caused by running out of good data or compute, but also running out of obvious ways of improving them. It was easy to fix the fingers, but when it is all photorealistic you no longer can consistently improve the model. Didn’t people in AI say that we will no longer talk about hallucinations by 2024, and yet here we are at the end of 2025 with no useful agents or hallucinations being solved.