How come places like Boston Dynamics have been building robots for decades but nothing happens, meanwhile AI comes out of nowhere and leaps further than robotics in 6 months?
How come places like Boston Dynamics have been building robots for decades but nothing happens, meanwhile AI comes out of nowhere and leaps further th...
because hardware and software are two entirely different fields you moron.
We have self driving lawn mowers now
What we've seen in with transformers has been in the making much longer than 6 months.
Happens all the time, technology keeps being stuck in the research phase until something clicks and a scalable approach is discovered.
We had that more than a decade ago. Making something that works in reproducible conditions that's not the hard part.
What's hard is making any sort of robot that's adaptable to any sort of situation. That's why you still can't buy a roomba that won't get stuck to date.
>We had that more than a decade ago.
They weren't self pilot. They were assist pilot.
>I don't see it so it doesn't exist
Dumb luddites
I don't know what you mean by assist pilot but lawnmower robots and roombas have been around for a long time.
During all that the tech hasn't fundamentally changed. They map the area once and drive around in a pattern.
Meanwhile you won't find any robot that you can just put somewhere and let it do it's thing as reliable as a worker tasked with it.
>self driving
you mean a glorified vacuum bot that just goes around in circles? Where are the fucking robotic maids?
I'm not glorifying it but we went from chatbots to chatgpt that can generate generic templates for anything you desire within a few months. Meanwhile I've been seeing those darpa retards kick their robots around for 20 years now and the best they can do is a fucking roomba?
>another day another simpleton who thinks AI is sentient intelligence and not a glorified text generator
out of nowhere? these algos are fifty years old kek
>How come places like Boston Dynamics have been building robots for decades but nothing happens, meanwhile AI comes out of nowhere and leaps further than robotics in 6 months?
Ever heard of DUMBs? Deep underground military labs.
They have advanced shit down there, rushing it out would make the normalfags seethe and want to /go back/, which goes contrary to ~~*their*~~ transhumanist vision.
Don't believe me? Look up Charles M Lieber on wikipedia, where he worked 2010-15, and what he inveted under "Nanoelectronics and the brain" and tell me if normalfags think this is even possible today. There's a study in the footnotes showing exactly how it's done lmfao.
They're lightyears ahead of the rest of us.
Now watch as the GPTnaggers jump me.
>everybody who laughs at my psychotic rants is a bot
whatever you say schizo
Meds, now.
>>>/x/
>AI comes out of nowhere
The first neural networks were created over 70 years ago
Please share a video of yourself generating an AI image 2 years ago even in the most basic of forms, I'll be waiting
Die
(You) generating them with a simple text command, not a fucking study about it, retard. The whole point of this thread is accessibility, I don't fucking care what technologies BD has if we're not benefitting from this as normal human beings.
>oh my science this photo has filters on it! IT's AI GENERATED!
pathetic
>deep learning-generated images from an open-source, fully reproducible repo is not deep learning because... it just isn't OK
the absolute state
>waaaah typing "python do_it.py" is too haaaaard
https://archive.4plebs.org/x/thread/30002259/#q30002259
1 year and 7 months ago
You're fucking black
Here's 2016, retarded consooomer
Deep learning was a thing since the 1950's and there has been no improvements since 2015. The illusion of progress you see is two-fold: 1- retardscaling of old tech and 2- previous tech, though public and even open source, not being presented in a web service easy for you as a retarded tech illiterate consoooomer to consooooom.
Robotics is hard. AI is fucking meme.
Another problem is that they're behind the times in a lot of ways. For example, they still typically use old-school path planning when deep learning-driven path planning has blown it out the water since the early 2000's. It's a similar situation to how computer vision waited until alexnet to adopt deep learning when it was already being destroyed by it since the late 90's.
>so bro why hasn't robotics advanced as fast as scaled up text predictors (now with special hardware)
it's a mystery