[ 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: 1.81 MB, 2160x936, 1294971353915.jpg [View same] [iqdb] [saucenao] [google]
2363452 No.2363452 [Reply] [Original]

Excuse me my stupidity, but could someone sum up for me why is it hard to make AI who is self-aware and has got human-like learning capabilities ?

>> No.2363453

We don't understand human learning capabilities

>> No.2363457

>>2363453
What about the self-aware part ?

>> No.2363461

Not enough processing power is available yet.

Just because your computer can calculate 98324 x 23948 in an instant, doesn't mean it is smarter than you.

>> No.2363465

(a) We don't know exactly how human learning works. There are various ideas on how to simulate it, including symbol manipulation, neural net classification, and so on, but none are exact.
(b) To evolve an intelligence from scratch takes way too much time. So we need the insights from (a).

But you can argue that some AIs are self-aware (there is self-modifying code... and look at real-time virus scanners), and that some have learning capabilities (mostly in terms of playing games like chess or go). They're just not human, and are limited to their environment.

>> No.2363468

>>2363457
>self-aware part ?

No exacty sure how self aware works either

>> No.2363471

>>2363465
>>2363461
Thank you. This all sums it up pretty well. Rest of the thread can be about AI in general as far as I care.

>> No.2363472

>>2363457
If we don't understand human learning capabilities, we CERTAINLY don't understand self-awareness.

>> No.2363477

we understand two thirds of fuck all about human consciousness,

hmmmm that pretty much sums it up really

>> No.2363492

>>2363477
and eloquently too.