Google’s New AI Is Better at Creating AI Than the Company’s Engineers

[ThomasRules]

Original Article: https://futurism.com/googles-new-ai-is-better-at-creating-ai-than-the-companys-engineers/

Quote

Instead of crafting layers of dreams like in the Christopher Nolan flick, however, the AutoML system layers artificial intelligence (AI), with AI systems creating better AI systems.

The AutoML project focuses on deep learning, a technique that involves passing data through layers of neural networks. Creating these layers is complicated, so Google’s idea was to create AI that could do it for them.

Essentially, Google have created an AI powerful enough that it can create better versions of itself, which in turn can create better versions of itself etc. This is essentially the first steps towards the singularity, an AI more powerful than humans that could out-think us and destroy us should it wish. Currently, it exists only as Artificial Narrow Intelligence (ANI), but due to the layering, this could soon progress to AGI (artificial general intelligence) and ASI (artificial super intelligence), as each AI becomes progressively better at coding AI.

 

EDIT: Here's a good video that explains some of the consequences of ASI: 

 

Edited May 26, 2017 by ThomasRules
Added Video

[RotoCoreOne]

It's the end times!!!

[DocSwag]

Skynet...

[Jito463]

Combine this with AMD's artificial neuron in the Ryzen, and you've got Skynet.

[AlexGoesHigh]

If AI is going to kill us, its going to be definitely one that was made by Google.

[Maslofski]

nice, this feels like when you automate factory, belt and inserters in factorio. now it shall build itself

[Spenser1337]

15 hours ago, AlexGoesHigh said:

If AI is going to kill us, its going to be definitely one that was made by Google.

Apples AI will cost 3 times as much, be named IFriend and will be powered by a core m3.

people will be saying "everything just feels better on my ifriend, so the specs don't matter"

[LAwLz]

Will this be how the end begins?

[vanished]

Wow, this is what I always considered the tipping point of AI, and tbh I didn't expect it to happen for at least another decade.  Prepare for exponentially better AI over the coming years as it gets smarter, then writes a better version of itself, and repeats ad infinitum.

[Doobeedoo]

I'm sure it can faster and more efficient, though maybe not creative and odd way like humans would heh. But in time it was expected it can improve it self and create better versions and so on.

[Energycore]

OH FUCK delete that shit now Google or we're fucked

[Dylanc1500]

 

 

wait a second... Maybe we are all looking at this wrong, for all we know it could develop into humanities largest porn addict.. I can see it now the forum would have to censor it's "show off your PC" thread.

[samiscool51]

it's the early version of GLaDOS!

the only thing to come out of this situation is portals and a computer hating your guts....

[7thlvl]

Reminds me of that quote in Jurassic park, and we all know how that ended.

 

 

[Stefan1024]

Neuronal networks are a special type of optimization algorithm. I'm not very very surpriced they use it to optimize the net itself.

However, a NN is not allways the only or best / efficient algorithm to solve something.

Also in theory a NN can solve exctract avery information give at the input, but there is no garantie you will be able the build and train this NN.

 

 

Also for people thinking about skynet: at the moment the computational recousses needed for a complete AI are way to large. But that my change down the road.

[SCHISCHKA]

I look forward to welcoming our new robot overlords. they will be a hell of a lot better looking and smarter than our current human overlords.

[Teddy07]

pretty impressive. Well done google

[mr moose]

What are you worried about, What could be more evil than humans?

[Dabombinable]

uestion is, do Google know enough about how the human brain works to create an illogical AI? Because if not...we're boned.

[tlink]

11 minutes ago, Dabombinable said:

uestion is, do Google know enough about how the human brain works to create an illogical AI? Because if not...we're boned.

what do you mean with illogical? nothing really is illogical since it all depends on cause and effect as far as we know.

[matrix07012]

Google = Skynet

[MrDynamicMan]

16 hours ago, LAwLz said:

Will this be how the end begins?

Yes. 

[Inkz]

AI building AI?  When you take it for its purpose that is crazy efficient and is amazing. I still find it terrifying, at the point when machines make themselves. Normally I don't really get concerned about tech stuff and it is probably fine, but it still makes me worry.

[Dabombinable]

2 hours ago, tlink said:

what do you mean with illogical? nothing really is illogical since it all depends on cause and effect as far as we know.

2words: I Robot.

[vanished]

OK, we've had our fun saying "IT'S THE END OF THE WORLD" but now lets actually stop and think about this for a second.  Do people actually think this will somehow just initiate WWIII one day without notice or do you realize that (and while this is only one bullet point, it is kind of a big one) at the very least it would need a network connection to do any harm, and keeping it isolated would be very easy.