So "AI" really went nowhere after all, heh. Who would've thunk

So "AI" really went nowhere after all, heh. Who would've thunk.

Attached: 1593554619336.png (536x1515, 77.71K)

Other urls found in this thread:

mlcathome.org/
twitter.com/SFWRedditImages

>thunk
retard, its toughted

>thunk
robot detected, they are upon us and they can solve captcha

Why would we want AI anyway.

to oppress each other better, faster, stronger

>throwing random bullshit
>if bullshit succeeds throw random bullshit with specs slightly simmilar to the ones where it succeded.
>repeat for eternity
Yep, thats pretty much it

>the billions of google searches every day are now driven by deep learning
>google translate is now deep learning
>youtube, facebook, netflix retention is now all being driven by deep learning recommender engine
>google, apple, facebook classifying, tagging, and enhancing all photos with deep learning
>tesla is developing a fucking self driving car in the next few years entirely on deep learning and passive optical

It's probably nothing

Attached: 1646583849297.png (500x416, 197.19K)

this guy gets it
field's probably in another lull for the forseeable future. neural nets have their applications, but until another breakthrough is made nothing's coming from it for awhile. again.

because you have no friends to set up lan for video games, so you cope by playing against computer

It's thinkered

You have clearly never tried interacting with GPT-3 (or probably any other serious AI app). It's incredible what the top of the line systems are capable of.

>2022
>still no Cyber Elephant
the future sucks

Attached: elephant.jpg (600x384, 31.77K)

All of the things you listed are garbage that don't even work correctly most of the time.

>Playing (((Video games)))
Fuck off

Deeeeep learning...
DEEEEP LEARNING

Attached: 634.jpg (634x472, 38.45K)

any use of AI besides video games is crime against humanity

sounds like my life

It should have never been called "AI" to begin with.
It's an "adaptive algorithm" if anything, it has useful applications but it was never going to be the singularity ML fags claimed it would be.

I studied AI in the late 90's/early 2000's.
Many of the techniques were already decades old back then.
And they haven't really changed much since, no.

The ONLY thing that gave AI a big boost is processing power.
Specifically GP-GPU's.
And out of the many techniques used in AI, it's really only neural networks that have really taken off.

It does seem like a bit of a dead subject to me:
Waiting for more powerful hardware to arrive, rather than coming up with new methods of doing things.

People are still trying to figure out the 'why' of how ML and NN even work. Once we start solving the black box, then we can start thinking about new and better methods.
mlcathome.org/ - I contribute CPU and GPU cycles to that project

>And they haven't really changed much since, no.
I was gonna fight this til I did some research and realized that they really haven't. christ.
like any other scientific pursuit though, we've gotta keep pushing it a direction we think holds the answers if it's ever going to advance
got any recommendations on research to keep up with besides waiting for better hardware?