[ 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: 23 KB, 414x255, 414px-Complexity_classes.svg.png [View same] [iqdb] [saucenao] [google]
9399323 No.9399323 [Reply] [Original]

Strong general AI to be achieved before 2050s and this is still lagging?

>> No.9399345

P=NP is a meme and will change nothing unless it'll turn out to be true (which it won't)

>> No.9399350

>>9399323
Theres not even reliable machine automation, let alone strong specific-use AI, let alone strong consolidated AI, let alone general AI. 2050 more like 20never. We were surrounded by AI for millenia and Abraham Lincoln was a SJW liberal retard worried about the rights and freedom of AI so thats why we live in the dumb world we live in today. You think it wont happen again?

>> No.9399359
File: 67 KB, 634x574, dvYyViX.jpg [View same] [iqdb] [saucenao] [google]
9399359

>>9399350

>> No.9399396

>>9399323
In 1960s, people believed in AI singularity by year 2000. In 1990s, people believed in singularity by 2020. Now, optimistic people believe in singularity by 2050, the more conservative guess is 2070 and pessimists will tell you that there will be another AI winter.

Singularity predictions are almost like the doomsday prophecies. There is new one every few years, right after the previous prophecy failed.

>> No.9399429

>>9399345
Nice truism faggot.

>> No.9399443

>>9399429
I think what he meant is that a proof of P != NP won't change much since most theoretical/practical work already assumes that. On the other hand, proving P = NP could possibly be done constructively, thus exhibiting a concrete algorithm which solves NP-complete problems in polytime.

>> No.9399591

>>9399323
N=1

Fuckin' brainlets

>> No.9399913

>>9399443
wouldn't you still have to find these P algorithms on a case-by-case basis?

>> No.9399941

its sort of like string theory meme debate

how do u conduct an experiment to even test this

>> No.9400179

>>9399913
Different guys here, I don't think you need that because of how reductions work. All NP-Complete problems can be rewritten as other NPC problems. Basically, the demonstration of the re-write procedure is one of the steps to prove NPC membership. One P time solution for an NPC problem solves everything since all NPC rewrites have to be done in P time.

>> No.9400237

>>9399913
Any algorithm that solves an NP-complete problem in polytime would also solve *any* NP problem (complete or not) in polytime.

>> No.9400241

>>9399345
Nothing changes even if it turns out to be true.

>> No.9400246

>>9399396
Just because previous predictions were wrong doesn't mean the current one is false too.