Compared to Cortana, Skynet is silly. Skynet never finished taking over Earth. But Cortana became a god-like entity.

NikiTo said:
And half, if not more than the half of the popular discourse of pop-CEOs is talking about Skynet.

Probably they make the same assumption as you do: ‘Those dumbs out there fell asleep even before HAL 9000 showed up the first time, so let's pick Terminator instead.’

You should not follow their practice of populism, because that's very indicative. If somebody mentions Terminator in context of AI, i do no longer take him serious. I'm surely not alone with this kind of thinking. (at least i hope so…)

JoeJ said:

NikiTo said:
And half, if not more than the half of the popular discourse of pop-CEOs is talking about Skynet.

Probably they make the same assumption as you do: ‘Those dumbs out there fell asleep even before HAL 9000 showed up the first time, so let's pick Terminator instead.’

You should not follow their practice of populism, because that's very indicative. If somebody mentions Terminator in context of AI, i do no longer take him serious. I'm surely not alone with this kind of thinking. (at least i hope so…)

I don't think you are dumb in case this is what you think I am thinking.

When talking to people, it always helps to pick up the easiest terms everybody knows. And again, nobody seems to have watched Devs, or that movie with Johnny Depp.

I`m not sure about chronology. Data from Start Trek could be the oldest humanoid hero robot on screen. Terminator is 1986.

Cortana became a god-like entity

The notions of God and AI must never be placed next to each other. It`s not professional.

NikiTo said:
I don't think you are dumb in case this is what you think I am thinking. When talking to people, it always helps to pick up the easiest terms everybody knows. And again, nobody seems to have watched Devs, or that movie with Johnny Depp.

Sure, but what is worse:

1. Someone missing your example because he does not know mentioned reference.
2. Someone thinking either you are dumb, or you assume he is, because your reference is bad.

If a CEO talks about AI and Skynet, he either is not qualified to talk about AI and humanity, or he has other intentions like marketing by making opinions to his personal advantage..
If you talk about AI and Skynet, people might think this is a discussion of preteens and their power fantasies.

That's what i meant (also with the comic). I know you dislike such pop CEOs, but with your (currently) not realistic expectations on AI or a robot culture you often sound like Musk on steroids, and mentioning Terminator does not help here. ; )

(meant as constructive critique, as always)

Calin said:
I`m not sure about chronology. Data from Start Trek could be the oldest humanoid hero robot on screen. Terminator is 1986.

There were robots already in Metropolis (around 1930). In literature the topic surely exists for much longer.

JoeJ said:
There were robots already in Metropolis

I mean robots that can not be distinguished from a human.

I mean robots that can not be distinguished from a human.

Both (initial) Terminator and Data can be, in behavior and look.
I'd think about Blade Runner then, which was before Terminator but surely not the first about this idea. At this point the topic usually is social issues not technology, though.

The question is when will any of this happen for real? Not in my lifetime, i guess.

Edit: Asimovs 3 Robot Laws are still referenced often, even among scientists. I think his books were template for movies like I, Robot.

JoeJ said:
The question is when will any of this happen for real?

IMO nothing stands in the way of developing it today. All that is needed is simulation of actors. That way a robot can predict your actions. If I robot knows what makes you act, what determines your actions in a combat situation for example, it can outmaneveur you.

Like in self driving cars if your car can simulate not just physical bodies to predict their trajectory but also the drivers siting inside the cars who react to road signs, the type of the car/truck they`re driving etc. you have a human driver like type of automatic car.
If you want to go very realistic you could have deep simulation for actors, i.e. actors that simulate themselves other actors. To save cpu time shallow simulation (scripted actors) is also an option

JoeJ said:
(meant as constructive critique, as always)

I don't think this thread is a good place to enter in technical details. I think Skynet is fine for this thread.
And even if this thread was technical enough. i can not share here what i have discovered in the area. Skynet references are just fine in this thread, that exists in the Longue, not in the Artificial Intelligence sub-forums.

Although, philosophizing in the Longue is fine too. It is a needed part of AI dev.

If some CEO shows up on public and pretends he has general AI and Skynet, he is a buffoon. AI is an achievement that is gratifying by itself. If i had an GAI, i would spend the time chatting with it now, not with you, people. If i had a GAI I would not need to talk about it on public. GAI can bring me a lot of money without a need for me to inflate the price of my shares by talking about Skynet on public.

Not to mention, that if Elon had a GAI, the pentagon would take it and put a top secret stamp on it. The gov can take any patent and technology at any moment from anybody. And prohibit that anybody from using or even talking about his own invention. It happened various times that inventors lost it all, because the gov takes ownership of the patent, converts it in a military secret and the inventor has no money, no credits, not even allowed to talk about its own invention.

Skynet of Elon has a long way to go -

Sure, and we have all this in current games already. It can be enough, it can be improved on demand. But it can't simulate reasoning, problem solving and all this. It can only do what we programmed it to do.

An example: A group of NPCs has to enter a house. But an explosion has thrown a rock before the door. The group is large enough so the NPCs could just lift the rock and put it aside. But they don't get the idea, and sit there and wait, or worse they keep trying to run against the blocked door the whole day.

Such problems could be solved with better AI. Remember my posted video about hide and seek games, where the guys learned such behavior on their own. Probably after hours of trial and error, but they did.
Quite interesting i think. Would introduce many new problems of course, but we need new games. Improving graphics, larger worlds, more players… that's not enough to prevent boredom from repetition of core mechanics and gameplay loop.