What are Veeky Forums's opinion on him and his claims of rat level general intelligence AI by the end of the year?

What are Veeky Forums's opinion on him and his claims of rat level general intelligence AI by the end of the year?

Absolutely based

Don't really have any reason to doubt him considering the success of alpha Go

what do you think the implications of an AI as intelligent as a rat? I think it's the beginning of an exponential progression towards human level AI. I think he's right when he says we'll have an AI as smart as a human by the end of the decade.

>I think he's right when he says we'll have an AI as smart as a human by the end of the decade.

Yep and fusion is only 20 years away. We live in magic times.

Overhyped and attracts a lot of autistic people, like OP.

Demis Hassabis is a fucking genius

It would definitely be a massive.

The transition to human AI would depend how much faith in Kurzweil's calculations and the rate of computational power growth. I think Exascale computing would be needed.

In what way is he overhyped?

>Human level AI in less that 4 years
>Not overhyped

gg user.

People also said we wouldn't have an AI capable of beating a Go master for the next 20 to 50 years yet here we are.

He should go back to making video games.

"People" say the earth is flat and that vaccines cause autism. That the objections of some arbitrary subset of individuals were overturned in the past says nothing about their future endeavors. I don't understand why people think this is a valid argument.

Because it was the opinion of experts

> "Experts"
I work in AI ( machine vision) and most "experts" Joe Public know about are cranks

It was the opinion of experts in the '60s that AGI was a decade away (at most).

Experts are about as good at predicting advances in AI as some random guy off the street.

>Andrew Ng, Geoff Hinton, and Yann LeCun are cranks

>rat level general intelligence AI

Total horseshit. That require total biological rendering of intelligence. They might have the same number of neurons in their model, but it will in no way be as intelligent as a stupid rat.

>They might have the same number of neurons in their model, but it will in no way be as intelligent as a stupid rat.

He literally said that the goal is something that can do everything a rat can

What's the go rank of a typical rat?

Honestly, the only thing they did was joining more and more neurons together in different patterns. There is no science behind any of it.

To accomplish such wouldn't even require the raw processing power of a rat.

transhumanism is real

I think his estimates are aggressive though

But what if the ending of the movie HER happens.

Celera BTFO