He thinks he can reach human-level AI by 2019

>He thinks he can reach human-level AI by 2019

will he do it bros?

Other urls found in this thread:

youtube.com/watch?v=rVlhMGQgDkY
en.wikipedia.org/wiki/Demis_Hassabis
twitter.com/SFWRedditImages

serious question: what the fuck is human level ai

We want to make our computers dumber so that our computers also need to invent computers so that they can do simple calculations like the prime factorization of 348398293982.

But obviously, we want our computers to also take 100 thousand years to build their computer-computers.

Or at least that is what I think. That is literally human level intelligence.

In other words: OP is retarded.

That post made no sense. Gtfo.

OP is talking about human level AI right?

Well, that is human level intelligence. Literally

>inb4 muh sungooolareeeteee memes

>OP thinks making the same thread over and over again is a good idea

will he kill himself bros?

No

>intelligence developing with complete access to all the knowledge of humanity would develop as quickly as humans that had to figure everything out on their own
I hope for your sake you're only pretending to be this retarded

It's like normal AI but dumber.

That's a fairly probably date.

Honestly I wouldn't be surprised if it was within 5 years.

How do you figure?

Because it's not that complicated. A lot of the human brain is dedicated to non-abstract thinking. Such as controlling the human body's motion and calculating movement.

Think of the difference between a chimp brain and human brain. It's not that large but we have very different abilities to problem solve and pattern match.

If you have the immense amount of money and top talent on the problem such that exists now in AI. It's only a matter of time.

who is this guy? what's his field of research?

>the more stagnant AI becomes the grander AI researchers claim become
lamo
come back when you can make robots that don't spill their spaghettis all over the place

ok, I'm back
youtube.com/watch?v=rVlhMGQgDkY

AI has advanced more in the last 5 years than the last 50 before it

It's actually pathetic to be on Veeky Forums and not know
en.wikipedia.org/wiki/Demis_Hassabis

It's just more meaningless pop-sci buzzwords.

you are fundamentally stupid and obviously low IQ to make such a statement.

There is zero link between replicating biological muscles and creating artificial intelligence.

You seem to be retarded to think that somehow "intelligence" is related to walking like a human. See: dolphins and octopus.

You basically are so stupid that you think watching a robot walk around is somehow important to AI.

Human level implies general intelligence that can adapt to new situations and problems while applying lessons learned on different problem sets.

It basically means you can ask it to learn to drive and it will adapt and learn to drive with minimal inputs.

It doesn't mean it will have a personality, motivations, or things like that though.

The number of "2s" in the current year has increased more in the last 17 years than in the 100 years prior.

Geoff Hinton said it, my autistic friend

>lol nothing's real :^)

Www Lld

The AI field is just lying to everyone. CERN is an evil athiest plot.

>when he's moving that box away from him and he's hopelessly trying to get it back

too kawaii 4 me

There are still a lot of things humans can do that AI can't.
Like watch a video and give a brief description of it.

>Like watch a video and give a brief description of it.

Geoff Hinton expects that by 2020

>He thinks he can reach human-level AI by 2019

source? I thought he said within ten years in 2014-2015.

Yeah I'm bullshitting, every time I make this thread I lower the date by a few years just to piss people off. What he really said was (I'm paraphrasing) "By 2030, we think we'll have solved AI for everything. And we think we'll be able to scale it up to superintelligence shortly after".

>
>It's just more meaningless pop-sci buzzwords.

You just revealed how little you know about AI.