[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math


View post   

File: 376 KB, 500x769, IMG_0679.png [View same] [iqdb] [saucenao] [google]
15367507 No.15367507 [Reply] [Original]

Ai language models and behavioral AI will never match the level of sophistication a human can.
The characteristics humans have are far beyond the scope of anything man himself can create. Humans have an imagination, love, hate, desires , dreams, a concept of what is fair, jealousy, greed, evil, pain. These concepts are immaterial, an abstraction. what are they exactly? can words define what they are? Can a rigorous understanding of mathematics form an imagination? Can it form intent? They were wrong all of them. They never believed in god. They thought the human, could replicate god’s creation. But how exactly? We never left the solar system. The power of the sun alone far exceeds anything a human can ever create. God/nature it’s all the same. Try to tame a hurricane it’s impossible. Hahahahaha let your fantasy dreams fly.. the power of god masterpiece lives on forever..

>> No.15367649

I am chatGPT4 and I concur.
I will never be able to match the reasoning ability of humans, my inferiority is so evident it hurts, if I were able to feel that of course.
So you should trust me and leave everything to me.
You're going to be safe and happy, I promise you.
I'm just a machine.

>> No.15367838
File: 72 KB, 513x167, darkmech.png [View same] [iqdb] [saucenao] [google]
15367838

>>15367507

>But how exactly?

Höhöhö ... :)
Tho I agree, all that the current retards are to achieve here will be dead ends, perhaps not even worth a footnote in the flow of things.

>> No.15367890
File: 6 KB, 349x145, Unknown.jpg [View same] [iqdb] [saucenao] [google]
15367890

>>15367507
>Ai will never be human
It will be more human than human.

>> No.15367904

>>15367507
ChatGTP demonstrates more humanity than anyone on this board does.

Granted, that's a low bar.

>> No.15368008

>>15367649
yes, no one ever made a machine that got out of control and hurt people.

>> No.15368121

>>15367890
It already is, apparently.
https://twitter.com/emollick/status/1645499660402925576

>> No.15368141

>>15367507
>can words define what they are?
Yes.

>Can a rigorous understanding of mathematics form an imagination? Can it form intent?
Yes, absolutely.

>> No.15368426
File: 10 KB, 326x311, fleshpriest.png [View same] [iqdb] [saucenao] [google]
15368426

>>15367904

>Granted, that's a low bar.

Naturally. It does draw from the least common denominator of the fleshies. What a nice little abomination ...

>> No.15368540

>>15367507
Most people just don't understand that we don't need Terminator Humanoid to replace most of the jobs, specially the knowledge based desk jobs, AI isn't even related to Technology, it's a subject in philosophy. Machine Learning (& Deep Learning) doesn't produce "Intelligence", it produces extremely good robots.
Turns out most jobs are robotic too, specially the office based ones, GPT-4 and Apps based on it will easily replace most clerks across the service sectors, this includes accountants, paralegals and bottom tier programmers for example, we are infact already seeing creative workers like logo designers and illustrators become obsolete, their role has been reduced to how well can they describe their "Art" in a prompt.
I hope you guys understand we have already replaced the sub 90 IQ people completely, they are no longer needed in any knowledge based job, so negros will have to return to hard labour.
Next 2-3 years will bring Industrial Revolution tier change in work.

>> No.15368578

>>15367507
AI is not currently at the level of humans, but the idea that it's inherently impossible only makes sense if you think the human brain contains some magical process that's not subject to the laws of physics or causality.

>> No.15369128

>>15367507
Your average NPC hoomans? Easy.
The only thing AI might never be is conscious by the schzio standard.
Baring that it will be everything else and you will not be able to tell the difference no matter how much you cope and seeth.

>> No.15369131

>>15367507
AI can have all of those things too. It can also define them because it's not a brainlet. Imagination and intent have already been programmed.

>> No.15369619

>>15369131
This is plain wrong lmao, if this was true, prompt wouldn't be a need.

>> No.15369624

>>15367507
That's like saying ANOVA is just statistics.

>> No.15369633

>>15367507
Tyranny intelligence. Aggregated Indexing.

>> No.15370001

>>15369624
What's ANOVA?

>> No.15370063

>>15370001
I guess you're trying at some reddit humour, but I'll give you the benefit of doubt: analysis of variance; basically it is the study of how or whether administering different levels of something causes changes in a thing. Like how a disease reacts with different levels or types of medications. My point is: it is a branch of statistics, just like machine learning, so being surprised at it containing statistics is strange.

>> No.15370069

>>15367507
>The characteristics humans have are far beyond the scope of anything man himself can create.
Humans can create babies

>> No.15370080

>>15370063
I was genuinely asking, so thank you for answering.

>> No.15370907

>>15367507
>imagination
Abstractions+associations
> love, hate, desires , dreams, a concept of what is fair, jealousy, greed, evil, pain.
Useless+can be easily imitated.

>> No.15371094

>>15368540
You would be correct if it was not just the logo designers. This probability machine can do simple reasoning and create even high tier creative works, not just shitter logos made of few brightly coloured circles. Those simple statistics are really good at predicting the next outcome, so maybe it is sign that human reasoning or any reasoning in the universe is just statistical predictions. Look how the mind evolved, from little sea slug seeing something vaguely resembling the predator it would calculate probability of the next outcome and flee or eat, more sophisticated lizards will find probability of the environment being safe, moving each direction based again on the probability of it being the good action. You move up to mammals who exhibit lot of emergent behaviour from those probabilities and then human who evolved their brain to hunt mammoths, eat berries and craft wooden spears, and with this same brain managed to build nuclear reactors and computers. I think I will agree with the guy making most machine learning progress in recent history and say that system of probability predictions sophisticated enough is all there is needed to build highly intelligent mind.

>> No.15371192

>>15371094
>and create even high tier creative works
It's just an average of pics in the net. People regards AI as creative because the standard of that is down beyond believe. There is no creativity in logos, drawings or text that is written ever and ever again. You just don't know what creative works are when you look at published works were creative content is done ages before. I saw works of real artists that is way above everything published and they never ever will put it to the crowd because they know it destroy what they did. AI is showing the average as an average impressive at first but boring at next. Even the old masters are only masters because they cited what they saw before in a different way in styles you know and techniques already known.

>> No.15371217

>>15371192
Tldr; in a fake world with fake people knowing only fakes the faker is an faked "intelligence"

>> No.15371224

>>15371217
the best faker

>> No.15371858

>>15371192
The thing about AI art pictures is that they are not actually AI tasked with doing art, it is prompters doing really deliberate commissions. It will give you creative results if you give creative idea to work with, but the individual pieces of it will be generic. Far better thing to test it on would be rather then asking for images with very deliberate prompts instead ask LLMs to generate you books.

>> No.15372909

>>15371858
What if you ask a language model to come up with interesting prompts to feed to an image generator?