Quantcast
[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

If you can see this message, the SSL certificate expiration has been fixed.
Become a Patron!

/sci/ - Science & Math


View post   

[ Toggle deleted replies ]
File: 225 KB, 775x583, 1503861304503.png [View same] [iqdb] [saucenao] [google] [report]
9289137 No.9289137 [Reply] [Original] [archived.moe]

Which one do you think would be better to ensure the continued survival of humanity?

AIs are most often depicted as singular collectives, but that's not necessarily the the only way an AI could exist.

There could be a singular all-powerful godlike AI, who would be like a messiah to humanity. This AI could potentially bring salvation to humanity for all eternity. But in case a crisis happens and something goes wrong, it would go wrong universally.

Or there could be multiple AIs varied by their own reasonings and aims, much like humans of today. These AIs would be superior to humans in almost all ways holding limitless possibilities within them. The variety should ensure humanity always have some of them as its allies, but it would also mean the possibility of constant dangerous criseses to face.

I have been thinking about this for some time (in the last half an hour under the shower) and can't really come to a conclusion. It feels like at the core it's boiling down to the question of monotheism vs. polytheism, but that might be a gross oversimplification. Same with fragile peace vs. constant war.
What's your take on this as the second smartest board on 4chan?


P.S.: Niggers are not considered as humans in this thought experiment.

>> No.9289163

>>9289137
>>>/x/

>> No.9289186

>>9289137
>dangerous criseses
Like mudslimes getting bombs?

>> No.9289190

>>9289186
More like SuperAIs getting nukes on an intergalactic level.

>> No.9289192

>>9289137
AGI would be able to interconnect with everything, no such thing as distinct entities.

>> No.9289196

You're misunderstanding AI as some sci-fi god-like figure. Don't worry. Your thought experiment is a total garbage that is not worth a discussion in first place. It's nothing more than a fucking function optimizer!!!

>> No.9289199

>>9289192
It could, but not everyone would.

Like how there are people without internet. And even those on the internet are still distinct entities, despite the groupthink it facilitates in some people.

>>9289196
And you are just a code monkey who never progressed above the 1st grade CompSci definitions.

>> No.9289205

>>9289190
You don't think of mudslimes getting bombs and blowing your ass up as dangerous?

>> No.9289206

>>9289205
It's just part and parcel of living in a diverse AI society.
Nothing unmanageable.

>> No.9289216

>>9289199
In big frame, thinking of the problem we solve, not the shinier tools we have, "AI research" has not changed much from 20 years ago when deep blue won against the chess grandmaster. This is not intelligence and we have no single clue nor made any progress toward the true "intelligence". I'm not underestimating the research field's efforts like unsupervised learning, reinforcement learning, imitation learning, etc, trying to make it happen. Surely neural network made a noticeable progress in machine learning these days. Yet, we never made a true progress toward "intelligence". Rather, we researchers just cursed by the word "artificial intelligence" to believe that what we research is intelligence, like how Google X wasted shit tons of time on its robot farm learning nothing. In fact, recent "AI researches" are either just application of neural networks, just as how some highschool students submit some "papers" to arxiv, or bringing math back to neural networks, realizing it's nothing more than a global function optimizer. Meanwhile, media and some famous scientist of other fields knowing nothing about AI shitting about how they threaten our life, is seriously harmful to researchers making them waste time on the illusions that just doesn't work. OP, your thought experiment is nothing new and it existed from at least more than 30 years ago like in blade runner with no real progress. How about doing some real research yourself instead of wasting your time on sci-fi?

>> No.9289228

>>9289137
>promethean AIs
>2017
Woah gramps the 70s called they want their sci-fi back

>> No.9289912

>>9289216
>How about doing some real research yourself instead of wasting your time on sci-fi?
I'm already doing in-depth work on improving the taste of my semen without relying on sugar intake. But being a man of many talents and a true polymath of our age, many fields catch my interest if I find them worthy of my time.

>> No.9290105
File: 15 KB, 768x309, joi.jpg [View same] [iqdb] [saucenao] [google] [report]
9290105

>>9289137
multiple AIs, for personal reasons

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
Captcha
Action