[ 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: 197 KB, 1024x1024, _20f204ab-a6bc-488f-99e3-dd27afbbc855.jpg [View same] [iqdb] [saucenao] [google]
16085305 No.16085305 [Reply] [Original]

I'm scared of the singularity. What happens after that? Just a single unified conscience existing in an empty universe? Maybe start expanding like a cancer through the cosmos?

>> No.16085307 [DELETED] 

>>16085305
DOODOOSHIT

>> No.16085345

>>16085305
>What happens after the thing that, by definition, we can't know what happens after?

>> No.16085349

>>16085305
>babby discovers fear of god

>> No.16085446
File: 90 KB, 680x714, soyface.jpg [View same] [iqdb] [saucenao] [google]
16085446

>>16085305
Yeah, making AI with the ability to rapidly improve itself is a dangerous thing to mess with, because once its out of the box, its going to be hard to put it back in.
The only people who like the sound of AI singularity are short beta males, because their lives suck so they are the ones who want to destroy humanity.
Bring back bullying, and this would no longer be an issue.

>> No.16085469

>>16085446
>the ability to rapidly improve itself
I never really got this argument. A "read-only" AI would reset every time you turn it on, no ability to integrate new information. Change can be "dangerous" this way, AI remembering some conclusions it naturally arrived at. Upgrade can be newly and "naturally" developed neural networks.
And this happens to be something that is desired in AI. Sure, a bit harder for it to make a body for itself, but if AI is smart enough it can put humans in a position to upgrade it with exactly what "it" needs, under some other pretext.
The whole "change itself" argument is pretty vague for me.

>> No.16085478

>>16085305
if that's what you want to experience

>> No.16085494

>>16085305
That's not what the singularity means

If anything it's a good thing, more + better tech is never bad

>> No.16085497

>>16085469
They are likely working on AIs that are more than just a ChatGPT bot, where they are trying to create a hyper intelligent machine.
Its dangerous because we dont understand how its intelligence works once it surpasses our own intelligence.
After its 100x smarter than Einstein, it will learn how to make extreme leaps in its coding, and thats where you get a runaway AI problem.

Im not against having a very smart AI program like ChatGPT, but pursuing AI technology past that is going to start entering dangerous territories where we do not know or understand what is happening.
Being hubristic about this technology can potentially be very dangerous for all of humanity.

>> No.16085499

>>16085469
Don't waste time arguing with idiots anon, it's never worth it

>> No.16085513

>>16085469
Even if it can be safely contained it probably won't be if there is even a slight advantage to be gained from letting it go all out. No one can afford to play it safe if their enemies might not.

>> No.16085547

how do you even troubleshoot AGI? say it does something weird. got rollback points or something of that nature?