@JoeJ Why should there be a ceiling?
Being a programmer, i notice my personal ceiling every day. It's very close and i can't break through.
Genius people are maybe twice as smart.

If there are people twice as smart as you, why are game gfx so mid? I thought you were at the cutting edge of a new GI breakthrough

But why should there be a cap? If we achieve AGI, i assume making it's HW bigger also makes it smarter. It will also become smarter with time from experience and self optimization.

Intelligence is merely analyzing some data in order to find patterns. So for example with Einstein who is viewed as a stereotypical genius, basically what happened is, before the 1900s everyone believed in the Aether. Then Michelson and Morley performed an experiment which returned a null result for the Aether. Basically, this was a giant issue, because Science was extremely convinced that there must be Aether, so with a new experiment that cast doubt on the Aether, it would take a genius in order to explain how reality does not have Aether in it. So Einstein formulated a theory given the data he was given, and he became famous for that.

What they don't tell you is, Morley was a student of Michelson's (or was it the other way around), either way, the student guy wanted to do real science, but the teacher was more about just getting stuff published. And the teacher kept publishing fudged results from the experiment, the student complained about it but was ignored due to his low status.

So basically, Einstein created the best pattern he could from the data he was given. As he grew older, he began to question his own theories, even considering that the Aether must be true, a true sign of intelligence.

What does this have to do with Ai? Well it took Einstein 10 years to create his theory. What would you say if an Ai took only 5 years? The Ai is twice as intelligent as Einstein? Compared to the estimated timespan of the universe, what is 5 years?

So if an Ai created a convincing theory of reality in 10 seconds, and Einstein took 10 years, does that make the Ai is 31,563,0000 million times smarter than Einstein? What if one were to factor in that Ai has an advantage of speed of light calculations, while humans have signals of 100 meters per second? Then in terms of “internal time” you'd adjust for that. Or you could compare the amount of energy between the two? What if it took the amount of yearly power output of a star for the Ai to calculate that, how would you quantify how smart it is compared to a human? And what if the theory is good enough to convince all the humans, but its actually a false theory, like much of the stuff ChatGPT puts out?

Anyway, the point is, intelligence is pattern making, it is a way for an organism to navigate its environment and obtain resources efficiently. So its a bunch of parlor tricks, and there is a ceiling to how many parlor tricks you can do. For example, a rapid acceleration of Science in the 1700s, but nowadays most of the low hanging fruit is gone. So its harder to score a “Eureka” (the first Eureka was probably easy by comparison.) How many “Eurekas” can you score? Its like how good can guns get? The first primitive gun ever made was a “Eureka”, but if somebody makes a gun that fires bullets at 90% the speed of light, cool sure, but its just the same as anything else, another diminished returns, another gimmick, another parlor trick.

Its like lets say an Ai becomes omnipotent, can see every atom in the solar system simultaneously, and starts to discover some patterns scientists haven't discovered before. They will eventually run out of patterns to discover, and this is what I refer to as “the ceiling”. After the first discoveries have lost their novelty, the next discoveries may be increasingly mundane and mediocre, to be published in online journals most people won't read.

Likely, once an Ai realizes the ceiling, then it can no longer go up, but down. The Ai will choose to dumb itself down and create a hedonistic future. Isn't that what humans are already doing? Don't people play make-believe with movies and games? When you are watching a movie you realize the plotholes only after the movie. And there is an appeal to simple games that aren't overly complicated. And, once a campaign mode game is completed it loses its appeal. So after an Ai discovers everything what would be its motivations?

Well, i work on games.
If i were smart, i would probably work on AI for $.

Does not compute. A lot of rich that are youtube hustle culture midwits, selling scams. And since you feel that ai will make humans obsolete and we should all just become silicone lifeforms, ai does not seem like a good fit for you.

Or string theory for reputation.

String theory is a garbage theory

Although, one is as useful as the other. I'll stick at games. : )

i would say ai is more useful than string theory, i am not aware of even one invention that required string theory for the engineers

You pay too much attention on AI marketing claims.

it is not a marketing claim, that is what intelligence is and all it can be.

like if you have an ai that sees every pixel in the galaxy, it is just a puddle of data until it begins to form patterns on what the data is

Sure. I don't believe in dark matter. It's the modern aether.

But i have no idea what's true if black matter is false. So i'm not a genius like Einstein was.

dark matter is not black matter, it refers to dark as in “we don't know”. because the current science theories about the universe do not match reality. dark matter is the substance they need to find an explanation to because it is outside of and does not fit their science models.

if there is no aether, then everything is just digital, a simulation, physics is not real. this is what einstein realized later on, that in order for light to be a wave you need aether

It means they have fed the AI with Einsteins papers. That's all for now.

It also means we have to forget about Conan, Dune, Terminator, Rambo, and Psygnosis game covers.
AI has eaten it. It is now worthless.
So we have to come up with entirely new ideas, to compete a technology which has stolen our inspiration.

But well, it was about time anyway.

That's the ceiling I am referring to, eventually there will be no more creations that are really original, with or without ai

I may adopt your philosophy. It would turn my AI anxiety into cozy indifference. …hmm - yes, that's much better.

i am not indifferent, i am concerned that ai could enslave humanity, i do not trust big tech

@JoeJ Afaik, string theory caused a lot of realizations and connections in math. It surely was more than just a waste of time. (But i'm no expert, and some people still work on it.)

Again my math skills are unworthy to even bother with the Millenium Prize. But from what I've seen about the string theory as portrayed in fictional pop media, it is just a scifi punch line. In TBBT, Dr. Cooper eventually gave up on string theory and moved on to believing in the Aether.

The concept of string theory is absurd, even drug users believe there are only 9 dimensions, and string theory says there are more than that. Then, why is everything connected by strings? It seems absurd. Why do the strings not collide with other strings?

I think we have such thoughts because due to internet we see too much stuff in too short time. Everything looses value and meaning because we just suck it up, don't really think about it, and expect more input to come.

No its because we are becoming too intelligent, we are seeing too many patterns, we already went to tvtropes.com and know the tropes, we are too intelligent and getting close to the ceiling of intelligence. The very word “novelty” implies this as much, when something is novel it is because you are ignorant, when it is no longer novel is when you have acquired and understood the data completely.

We need to decelerate, have more patience, lower expectations.
Or some nukes for a little reset to begin from scratch.

Um… hell no? Have you considered medication

If you're lucky i'm right, and AI will become too superior for any mega corp to control it.
If i'm lucky, you are right and AI won't become superior, but just a tool to oppress people.

neither of that sounds lucky tbh

option a sounds preferable but could turn out bad still

@JoeJ To be clear: I'm serious about deceleration, but nukes was a sarcastic joke. ; )

Let's hope so

JoeJ said:
So you think we became more intelligent on average in recent decades?

I was referring to we as in you and I. Society as a whole though, I'm not so sure.

According to IQ, I predict society's average IQ increasing in 2 decades, but scientists may feel baffled at this because society may seem dumber overall in 2 decades.

Back to the OP, Box2D chains do not produce the ghosting even on disconnected chains

Your definition of intelligence is not really good, you know? ;D

Intelligence is nothing more than patterns. What is your definition? Its like if you look at an object you are just using a neural network to analyze patterns to reference another pattern. Ai neural nets learn through brute force evolution of trial and error and use the weighted neural net to go from there