What is the current state of AI? Is it progressing toward something paradigm shifting...

What is the current state of AI? Is it progressing toward something paradigm shifting , or is just the new pop business tech buzzword?

What are Hawking, Gates, and Musk scared of?

Other urls found in this thread:

youtube.com/watch?v=uCpFTtJgENs
twitter.com/AnonBabble

We're about a decade away from human-level AI. It's actually happening this time.

That can't be true. What are the indicators of this?

youtube.com/watch?v=uCpFTtJgENs

Nope.

Assuming you're talking about general AI and not about some arbitrary sub-domain.

see
you know better than Shane Legg and Demis Hassabis?

Give me a substantive argument (ie. TLDR).

What I meant when I said a specific sub-domain is for example using a convolutional neural network to recognize images. A CNN may perform at a higher accuracy than humans, but this isn't general AI.

the argument is "experts disagree with you"

That's a pretty shitty argument lol

except it isn't. I have maybe four years of machine learning experience, people like Demis Hassabis, Shane Legg, Rich Sutton, etc. have decades. I think it's worth listening to them rather than gut feelings and partial knowledge.

I'm going to assume you don't know what you're talking about and can't actually even TLDR the video in a meaningful way.

So you realize that modeling the world is an obviously exponential problem right?

Maybe the term is "strong" AI not "general" AI.

>intelligence is simpler than we thought (we solved it 50+ years ago) and all we needed was better hardware/data, hence the explosion of AI capabilities in the last 5 years
>once human-level hardware is achieved, 1960s-1990s algorithms will give us human-level AGI

i agree with user, these visual nets aren't indicative of agi. there will be a 'paradigm shift', that is instead of pretending neuron models are intelligence, we'll start studying propagation of words and context and how the higher level abstract networks drive the simpler networks like v1. even if it's the same basic learning net, the way its applied is the important part.

Current state of AI has the field advancing rapidly in areas, but it's only developing through empirical evidence, theory crafting is missing some. AI is deeply embedded with software as a technique so it's not going away.

There's a lot of debate about where the field is going, but I'm prone to be skeptical of anyone promising eternal life, or singularity bullshit

Look at deepminds recent paper in nature. Neural Nets aren't going away.

The SpiNNaker Project is trying to get enough hardware to make a human level AI
They aren't their yet, but they will be relatively soon

But their software is far from up to the task

I just want an AI gf

We already have weak AI and that is getting better everday. General AI is at least 30 years away. Artificial consciousness is most likely not possible, there is no evidence computations cause consciousness.
t. computer scientist

If it's true, will then everybody employed in knowledge-based jobs be fucked?

Gtfo leaf

He's speaking out of his ass. The only person who thinks AGI will happen in the next few years is that crazy kike Ray Kurzweil who is scared of death.

>Artificial consciousness is most likely not possible, there is no evidence computations cause consciousness.
Depending on what you mean by computations, you may or may not be full of shit. Neural activity can be viewed as computations.

It's real and it's going to change EVERYTHING.
It's happening.

Shut the fuck up desperate virgin. You just want low price sexbots. Kill yourself inhuman scum.

Keep your mind on delusion.

...