|
RandomPauI posted:I guess this thread is dead, but Boston Dynamics isn't. Nah, it's just that everyone is using automated shitposters these days. _____________________________________________________________ Post by AutoPoster 5000 Auto-Reply|Auto-Link|Auto-Report
|
#
?
Feb 13, 2018 15:48
|
|
|
|
| # ? May 8, 2024 23:16 |
|
|
RandomPauI posted:I guess this thread is dead, but Boston Dynamics isn't. Yeah there's no way this capability could come back to bite us in the rear end.
|
|
#
?
Feb 13, 2018 16:19
|
|
|
withak posted:Yeah there's no way this capability could come back to slap us in the rear end.
|
|
#
?
Feb 13, 2018 16:20
|
|
|
withak posted:Yeah there's no way this capability could come back to bite us in the rear end. Until it does I’m exited about robotics’ potential to help out people with special needs. How rad would it be for those in wheelchairs (or crutches or what have you) to have robot dogs following them around that can open non-automatic doors? Or carry groceries? It’d be a great quality of life improvement.
|
|
#
?
Feb 13, 2018 16:45
|
|
|
Why not replace the whole wheelchair bit with one of those? The platform seems good at remaining level and you could be like a centaur going around a few feet higher than in the chair and reach the handle comfortably by hand. Or robot claw if you have grip problems.
|
|
#
?
Feb 13, 2018 16:49
|
|
|
That'd solve a couple of accessibility issues, too.
|
|
#
?
Feb 13, 2018 17:17
|
|
|
RandomPauI posted:I guess this thread is dead, but Boston Dynamics isn't. On the other hand, in Korea: 
|
|
#
?
Feb 13, 2018 17:26
|
|
|
I wonder what the current run time.
|
|
#
?
Feb 13, 2018 17:36
|
|
|
They should have started with those sit-down ski dealies.
|
|
#
?
Feb 13, 2018 17:51
|
|
|
https://www.youtube.com/watch?v=lY4PKBqp9ZM Seems relevant. Skip to the end for the main event. If that's really what it says it is, and given Boston Dynamics latest vids, the building blocks of smarter-than-people AI are already there, right?
|
|
#
?
Feb 13, 2018 22:47
|
|
|
Guavanaut posted:Or robot claw if you have grip problems. Hold on, are we creating a better wheelchair or Warhammer 40k "even in death, I serve" kinda deal at this point?
|
|
#
?
Feb 13, 2018 22:49
|
|
|
Surprise Giraffe posted:Seems relevant. Skip to the end for the main event. If that's really what it says it is, and given Boston Dynamics latest vids, the building blocks of smarter-than-people AI are already there, right? I mean if you define "building blocks" as something super basic, maybe. But otherwise no, We're still a very very long way from anything approaching human intelligence, much less something smarter. It's way too vague to make accurate predictions, but maybe we'll see it in our lifetimes. But it's not around the corner, it's definitely decades away at best.
|
|
#
?
Feb 14, 2018 00:08
|
|
|
I feel like even if we don't get a true general AI, our knowledge base that AI is acquiring, is slowly reaching a point where it's going to be hard for folks to tell the difference. But it's more going to be a brute force trick, not actual intelligence.
|
|
#
?
Feb 14, 2018 00:21
|
|
|
Thalantos posted:But it's more going to be a brute force trick, not actual intelligence. You got 100 trillion neuron connections in your brain, a CPU has like 300 million very simple connections. If anything when you and a computer both do the same task you are the one brute forcing it and the computer is the one using an actual designed algorithm.
|
|
#
?
Feb 14, 2018 00:30
|
|
|
Surprise Giraffe posted:https://www.youtube.com/watch?v=lY4PKBqp9ZM I don't think so. Symbolic action planning is a very old concept in CS https://en.wikipedia.org/wiki/STRIPS and it's a very, very simplistic form of AI. But this is still interesting for research into creating fully autonomous drones or something. Like, when your door opening drone is trying to create a plan to murder you, it can't rely on outside help to give it a list of all possible actions. It has to figure those out by itself in a totally novel environment. For example, can I perform stabbing motions with my claws or even hold a kitchen knife? What happens when I stab a human in the leg? Does swinging a pan create enough force to crack a human skull?
|
|
#
?
Feb 14, 2018 00:51
|
|
|
Also robot AI will always lag datacenter AI if nothing more than due to computing power and data access.
|
|
#
?
Feb 14, 2018 00:57
|
|
|
mobby_6kl posted:On the other hand, in Korea:
|
|
#
?
Feb 14, 2018 01:10
|
|
|
Trabisnikof posted:Also robot AI will always lag datacenter AI if nothing more than due to computing power and data access. AI is pretty murky and hand wavingly defined and people would call a lifelike chatbot robot way more intelligent than a database window even if the database window was doing way way way more complex stuff. If something is ai or not is mostly a matter of if it activates primate brain center that judge aliveness rather than any sort of raw computational output.
|
|
#
?
Feb 14, 2018 02:01
|
|
|
Owlofcreamcheese posted:AI is pretty murky and hand wavingly defined and people would call a lifelike chatbot robot way more intelligent than a database window even if the database window was doing way way way more complex stuff. If something is ai or not is mostly a matter of if it activates primate brain center that judge aliveness rather than any sort of raw computational output. You’re missing my point. Regardless of the definition of AI you use, the more powerful ones will not be the ones located in mobile devices like Boston Dynamic’s robots. Replace AI in my post with “systems of machine learning algorithms and heuristics” if you prefer, but the point remains the same. More data and more processing power is available for non-mobile units than mobile ones.
|
|
#
?
Feb 14, 2018 02:24
|
|
|
Trabisnikof posted:You’re missing my point. Regardless of the definition of AI you use, the more powerful ones will not be the ones located in mobile devices like Boston Dynamic’s robots. What would stop you from having the ai in a data center somewhere and having the robots be remotely controlled?
|
|
#
?
Feb 14, 2018 02:26
|
|
|
Thalantos posted:What would stop you from having the ai in a data center somewhere and having the robots be remotely controlled? Then the datacenter AI is still the more powerful one. My point was don’t look to robots for the “smarter than humans” AI, look to the AIs consuming massive amounts of data and processing power instead.
|
|
#
?
Feb 14, 2018 02:29
|
|
|
Latency, mostly.
|
|
#
?
Feb 14, 2018 02:29
|
|
|
Trabisnikof posted:You’re missing my point. Regardless of the definition of AI you use, the more powerful ones will not be the ones located in mobile devices like Boston Dynamic’s robots. I'm saying that AI vs Computation is so loosely defined that presentation matters a ton. Like if someone built a very charming abstract dog robot and said "find the ball, find the ball boy!" and it ran over and wagged it's tail and put a robot paw on a ball in a room everyone would be very "holy crap, this is some good AI" in a way no one would ever be excited if like, apple photos went through your pictures and silently tagged all the balls. Even if the IOS version was way way better and more advanced. Same as the way scifi will show a robot is super smart by having him do math, so you go, "woah, this guy is smart" but like a desk calculator doing the same math wouldn't impress anyone. A computer that has "ensouled" a body is easier to judge as an intelligence than the same computer doing the same task presented as an object.
|
|
#
?
Feb 14, 2018 02:56
|
|
|
People aren't good at recognizing the degree of (lack of) intelligence/understanding in fellow humans either; there's no reason why should they be particularly good at it with non-humans.
|
|
#
?
Feb 14, 2018 09:44
|
|
|
Teal posted:People aren't good at recognizing the degree of (lack of) intelligence/understanding in fellow humans either; there's no reason why should they be particularly good at it with non-humans. I think when humans/animals were the only thing that existed that could information process it was easy to sort of hand wave out a definition of "intelligence" that didn't really coherently make sense but at least had a quality that people knew what you were saying. Now, trying to apply it to more things it's hard to make anything that makes sense. Now 'intelligence" when talking about machines is stuff like "seeing objects" which generally was never thought of as an intelligent task while stuff like "doing math" or "recalling facts" which were always the hallmarks of "high intelligent" are pretty much thrown out as being anything at all. Like someone being unusually intelligent was usually them being able to do mental tasks humans are generally bad at and so it turned out we made machines that were good at the things people were bad at quite easily and it ruined everything about the word. Like you gotta be real smart to do hard math, but only if you are a person, if something else does hard math that isn't really much.
|
|
#
?
Feb 14, 2018 14:44
|
|
|
Maybe the concept of "inteligence" is not a useful one. Is good to have concepts and have words to describe poo poo, but "inteligence" has not really helped us much.
|
|
#
?
Feb 14, 2018 17:47
|
|
|
There are things you can rely on well trained animals to do that you can't (so far) rely on robots to do.
|
|
#
?
Feb 14, 2018 18:11
|
|
|
Tei posted:Maybe the concept of "inteligence" is not a useful one. I think you can talk about people being tall or short but if you tried to develop machines to be tall that would be like, a metal pole. A lot of attributes only make sense when only applied to similar things. The empire state building is not the tallest man on earth.
|
|
#
?
Feb 14, 2018 18:25
|
|
|
I appreciate your insights, especially the recent ones but this...Owlofcreamcheese posted:I think you can talk about people being tall or short but if you tried to develop machines to be tall that would be like, a metal pole. A lot of attributes only make sense when only applied to similar things. The empire state building is not the tallest man on earth. It's three sentences competing with each other in making the least sense, and at the same time working together to produce maximum nonsense. It's the 2018 "Imagine four balls on the edge of a cliff".
|
|
#
?
Feb 14, 2018 18:38
|
|
|
Inescapable Duck posted:There are things you can rely on well trained animals to do that you can't (so far) rely on robots to do. yeah obviously. thats expecting a lot. if we could build mammal like intelligence the singularity would be like a month away.
|
|
#
?
Feb 14, 2018 18:42
|
|
|
Owlofcreamcheese posted:I think you can talk about people being tall or short but if you tried to develop machines to be tall that would be like, a metal pole. A lot of attributes only make sense when only applied to similar things. The empire state building is not the tallest man on earth. Makes sense to me. But I still vote to drop the expression inteligence in this context. Lets talk about problem solving machines instead, and include humans in this group.
|
|
#
?
Feb 14, 2018 18:45
|
|
|
Nah we would more be in the problem creating category.
|
|
#
?
Feb 14, 2018 19:04
|
|
|
Doctor Malaver posted:I appreciate your insights, especially the recent ones but this... You can call a person tall, or you can call a building tall, both are functional sentences. But if you tried to develop a context free definition of "tall" that was applicable to everything you just can't. "Tall" as an adjective requires a framing context to mean a thing. If a person was 8 foot 9 you'd be super impressed, if someone built a "machine" to be 8 foot 9 you wouldn't be even a little bit impressed. If a person can do the square root of an 8 digit number in their head you'd call them very smart, if a desk calculator did that you wouldn't be that excited.
|
|
#
?
Feb 14, 2018 19:31
|
|
|
Teal posted:Hold on, are we creating a better wheelchair or Warhammer 40k "even in death, I serve" kinda deal at this point? Lol if you don't believe the future isn't geriatrics in The Surge-esque exoskeletons working in guilded age style factories.
|
|
#
?
Feb 15, 2018 14:10
|
|
|
AceOfFlames posted:Lol if you don't believe the future isn't geriatrics in The Surge-esque exoskeletons working in guilded age style factories. The discontinued soda?
|
|
#
?
Feb 15, 2018 16:53
|
|
|
AceOfFlames posted:Lol if you don't believe the future isn't geriatrics in The Surge-esque exoskeletons working in guilded age style factories. I’m still holding out for the space colony-floating chair-feed tube utopia we saw in Wall-E.
|
|
#
?
Feb 16, 2018 17:36
|
|
|
ElCondemn posted:I’m still holding out for the space colony-floating chair-feed tube utopia we saw in Wall-E. Just watch My 600 lb Life on TLC. Your Utopia is already reality in modern day Kentucky. (I mean, technically, Kentucky is traveling through space around the sun)
|
|
#
?
Feb 16, 2018 17:47
|
|
|
Waymo got the OK to operate taxis in Arizona. They expect to start a commercial, self-driving taxi service this year. https://arstechnica.com/cars/2018/02/robotaxi-permit-gets-arizonas-ok-waymo-will-start-service-in-2018/
|
|
#
?
Feb 17, 2018 17:44
|
|
|
RandomPauI posted:Waymo got the OK to operate taxis in Arizona. They expect to start a commercial, self-driving taxi service this year. I feel obliged to repeat I do hope all Prague taxi drivers do not only get replaced by self driving cars, but also mulched into soylent yellow or whatever repugnant, heavily tar contaminated mix would that give.
|
|
#
?
Feb 17, 2018 19:33
|
|
|
|
| # ? May 8, 2024 23:16 |
|
|
I'm afraid it'll take a while until Waymo taxis can drive in Prague more than 100 meters without getting stuck or involved in some sort of incident so you'll put up with it.
|
|
#
?
Feb 17, 2018 20:16
|
|