Elon Musk Unveils Prototype of Humanoid Optimus Robot (theverge.com) 75
Tesla CEO Elon Musk revealed a prototype of a humanoid robot that he said utilizes the company's AI software, as well as the sensors that power its advanced driver assist features. The Verge reports: The robot was showcased at Tesla's AI Day, and reps said it features the same technology used to enable the Full Self-Driving beta in Tesla's cars. According to Musk, it can do more than what has been shown, but "the first time it walked without a tether was tonight on stage." Musk said they're targeting a price of "probably less than $20,000." The back doors of the stage open to reveal a deconstructed Optimus that walked forward and did a "raise the roof" dance move. Musk would admit after the motion that they wanted to keep it safe and not make too many moves on stage and have it "fall flat on its face." "It'll be a fundamental transformation for civilization as we know it." said Musk.
Afterward, the company showed a few video clips of the robot doing other tasks like picking up boxes. Then Tesla's team brought out another prototype that has its body fully assembled but not fully functional. [...] Future applications could include cooking, gardening, or even "catgirl" sex partners, Musk has said, while also claiming that production could start as soon as next year. Musk says the robot is "the most important product development we're doing this year," predicting that it will have the potential to be "more significant than the vehicle business over time."
Musk first announced the "Tesla Bot" at last year's AI Day.
Afterward, the company showed a few video clips of the robot doing other tasks like picking up boxes. Then Tesla's team brought out another prototype that has its body fully assembled but not fully functional. [...] Future applications could include cooking, gardening, or even "catgirl" sex partners, Musk has said, while also claiming that production could start as soon as next year. Musk says the robot is "the most important product development we're doing this year," predicting that it will have the potential to be "more significant than the vehicle business over time."
Musk first announced the "Tesla Bot" at last year's AI Day.
Re: (Score:3)
If only there could be a single link to a video somewhere. Anywhere...
But that's not the job of slashdot, right?
Re: (Score:3)
FWIW: https://www.youtube.com/watch?... [youtube.com]
Re:Hmmm ... (Score:4, Informative)
That link was a lie stream of AI day.
Here's a video of the robot's walk: https://www.youtube.com/watch?... [youtube.com]
Re: (Score:3)
That link was a lie stream of AI day.
Here's a video of the robot's walk: https://www.youtube.com/watch?... [youtube.com]
The most impressive part was it picking up the watering can, but I think that's far from state of the art. It also carried the box... but might have crushed it a bit?
But I think the real problem is that he's trying to build a robot to perform human tasks, but the thing that humans are really good at isn't picking up watering cans, it's understanding that certain things are plants and knowing when they need to be watered.
All of these tasks are extremely unstructured and require a ridiculous amount of AI befo
Re: (Score:2)
https://say-can.github.io/ [github.io]
Re: (Score:3)
Re: (Score:2)
Yes, Musk is trying to solve the much larger problem of Artificial General Intelligence which goes far beyond programming robots to perform repetitive assembly line tasks.
It is the same problem he is trying to solve with automated cars. In fact, they used the same computer that they put in the cars and just trained it on a different data set.
Still a very difficult problem. I wish him luck. If they can crack this problem, it will be amazing.
Re: (Score:3)
Yes, Musk is trying to solve the much larger problem of Artificial General Intelligence which goes far beyond programming robots to perform repetitive assembly line tasks.
It is the same problem he is trying to solve with automated cars. In fact, they used the same computer that they put in the cars and just trained it on a different data set.
Still a very difficult problem. I wish him luck. If they can crack this problem, it will be amazing.
I'd be more optimistic if he didn't have a such a well documented history of massively underestimating the size of the problem (self-driving cars) and if his demo didn't focus so much on potentially mechanical tasks and nothing about how the robots even get their instructions.
Telsa is an awesome Electric Car company, and their solar + powerwall stuff sounds really cool, but they're only a decent, not great, AI company. And Musk's recent history (impulsive Tweeting, buying Twitter then having instant regret,
Re: (Score:2)
That link was a lie stream of AI day.
A little Freudian slip there lol.
Re: (Score:2)
LOL
Re: Hmmm ... (Score:2)
he will employ them (Score:3)
He tried to that (Score:3)
Re:he will employ them (Score:4, Interesting)
Robots doing assembly line work. Elon has plans.
Robots building habitats on Mars so that they're ready when humans get there.
Re: (Score:1)
Robots doing assembly line work. Elon has plans.
Robots building habitats on Mars so that they're ready when humans get there.
Another 'musketeer' circle jerk here, somebody bring the cold water hose.
tax robot labour at the minimum wage rate (Score:1)
This is the only way to ensure a benefit for humanity / civilization / conservatism / insert_own_option.
The tide lifts all boats. Not just the expensive ones.
Re: (Score:2)
I'd imagine that colonizing Mars is a lot easier when you have service robots doing the grunt work like preparing the food or taking out the trash.
All part of his master plan, I guess.
Re: (Score:2)
Ha (Score:1)
I'm still waiting for my waterproof cyber truck https://gizmodo.com/elon-musk-... [gizmodo.com]
Good.. (Score:5, Funny)
Time to get it to fight a boston dynamics robot to the death
Re:Good.. (Score:4, Funny)
Re:Good.. (Score:5, Interesting)
It certainly had that constipated look as it walked that is common with humanoid robots that can't quite balance fully upright. That just shows how limited AI is today, something a 2 year old can do is beyond the capability of cutting edge software.
The whole demo was smoke and mirrors, just like the "full self driving" ones from 2016 and 2018. Carefully controlled environment and pre-planned set of manoeuvres. Notice how the prototype that was able to walk didn't demonstrate any vision capability at all, for example. Their video did, but no evidence that the robot was actually using that vision to plan operations, or even that it was real and not some mock up the art department made.
Notice also how they didn't touch the robot. Boston Dynamics love to show off how their robots can cope with being shoved or bumped into, and deal with uneven terrain. None of that here.
Musk consistently over-estimates the capability of AI, and companies that have actual working robots and actual self driving cars only use it in very limited amounts to do things like object recognition. Speaking of which, a new flaw in Tesla's "full self driving" has come to light recently. Bikes with two tail lights seem to fool it into thinking they are a car far off in the distance, so it rams into them at high speed. That's why everyone else is still using radar.
Re: (Score:2)
Now that he revealed his politics you are completely critical
QUICK! I have no rebuttal to perfectly logical arguments. Let me switch over to politics that is a topic that I, myself, am introducing and blaming you for!
Like some people need to chill on the defaulting into politics thing. The person made perfectly reasonable arguments and not once mentioned politics. Just in case you were wondering why you're currently sitting at -1 for your comment's score, it's just straight up flamebait.
Re: (Score:2)
The whole demo was smoke and mirrors, just like the "full self driving" ones from 2016 and 2018. Carefully controlled environment and pre-planned set of manoeuvres.
If only the engineer said something like "We've got quite a long way to go. We just wanted to show you what we accomplished in 6 months." Oh wait. He did.
Re: (Score:2)
Except that Musk said Full Self Driving would be available in 2016, then 2017, the 2018, then Robotaxi would be ready for 2019 and Tesla cars would be appreciating assets, and then 2020, and finally in 2021 he admitted it might be 2024 before it's ready.
2024 would be 8 years after he first sold it to unsuspecting mugs, whose cars will by then be out of warranty on both the vehicle and the battery. He did promise them free upgrades to the hardware needed for Full Self Driving though, so their 8 year old vehi
Re: (Score:2)
Do you mean artificial death? Can an entity which is artificially alive truly die or only artificially?
Re: (Score:2)
Maybe artificial undeath?
Re: (Score:2)
the boston dynamics thing is real
stop acting like elon is telling the truth. he doesn't do that
Does he not have youtube? (Score:4, Interesting)
This doesn't appear to be doing anything close to what we have seen Boston Dynamics do for years now, which is fine, you have to start somewhere but the pomp and circumstance here feels a little unearned.
I mean credit to the engineers for getting a working prototype but I feel like the bar for humanoid robots is a little higher today.
Now the big thing is if it hits a $20K price tag, that's something for sure, but I can't see that happening anytime soon, can this thing justify what would be a multi billion dollar production facility to hit that kind of volume?
Re: (Score:2)
Re: (Score:2, Insightful)
I just took a look at a youtube video of a Honda Asimo. First debut was 2000, although a few prototypes existed before it the spacesuit looking model we all know and love today came out then. Price tag was steep at $2.5m. Even a Boston Dynamics spot is around $75k. You're right, $20k is a sweet price.
At that price, they suddenly become cheap enough to help lazy slobs like me at home. "Hey Optimus, get me a Dr. Pepper from the fridge" or "Hey Optimus, do a load of laundry" or "Hey Optimus, take out the
Re: (Score:3)
The intentions of the marketing are quite fantastical and seemingly completely divorced with feasibility.
Musk does appear to have some obsession with "catgirls" (sex robots) or at least he thinks that it's something that people really want to have and
Re: (Score:2)
Sexbots are just a way for companies to legally sell sexual services. They will be subscription based, with many DLC options. Their AI will be designed to be as addictive as possible, mainly targeted lonely guys.
As well as the exploitation angle, there are all sorts of ethical issues. Musk has already shown he doesn't care about what's right, with his public alpha test of his self driving vehicles already having claimed many victims. He's a billionaire so there aren't any consequences for him.
Re: (Score:2)
The only task of the three you mentioned that would be feasible for a current gen robot is getching a Dr pepper from the fridge, because it has a clear planning goal with rigid objects involved. Taking out a garbage bag which involves removing the bag from a bin - which is floppy, variable in size, weight and balance which may have a tight fit or vacuum would be a quite hard planning challenge for a robot.
Likewise, dealing with laundry is challenging because of the huge variety of shape and size, as well a
Re: (Score:2)
My expectations were much lower than what you percieved them as.
For laundry, I have laundry baskets as hampers. Chucking a hamper in one of those newer washers that don't have an agitator in the middle, then moving it to the dryer when the wash cycle is complete doesn't seem too tough. Emptying the dryer doesn't seem too tough. I wouldn't expect folded/ironed clothes at this stage of the game, just dirty hampers go out, clean ones come back. (Who irons in this age of wrinkle free polo's and khaki's?)
Tras
Re: (Score:1)
For the clothes scenario you have described, where you keep dirty laundry in a hamper and have wrinkle free clothing, there's practically a solution already - get a washer/dryer model that has an automatic washing liquid dispensers, and use the machine directly as your dirty clothes hamper. so all you have to do is close the door and press the button. A few hours later you have clean and dry washed clothes.
Re: (Score:2)
>can this thing justify what would be a multi billion dollar production facility to hit that kind of volume?
I think they want to know the answer to that question, and thus did exactly what they did. Now they can sit back and see how excited the internet gets about "cat girl robots" and have a better idea of the answer to that question.
Re: Does he not have youtube? (Score:2)
The dexterous hand is what takes it to a level above Boaton Dynamics. That said, in terms of agility, looks like Boston Dynamics Handle is ahead of Tesla. Still if they can make these generally available for under $20K like they claim, I all sure various startups can use it as a baseline to make better hands, feet etc. Once they get higher framrate cameras and better compute, this thing would be badass. If they keep iterating on it instead of resting on laurels or giving up. They should have a whole team ju
Re: (Score:2)
The dexterous hand is what takes it to a level above Boaton Dynamics. That said, in terms of agility, looks like Boston Dynamics Handle is ahead of Tesla. Still if they can make these generally available for under $20K like they claim, I all sure various startups can use it as a baseline to make better hands, feet etc. Once they get higher framrate cameras and better compute, this thing would be badass. If they keep iterating on it instead of resting on laurels or giving up. They should have a whole team just working on instrument ing and optimizing the hand, which is in my opinion the most important part. The foot as well.
Did you see dexterous hands? Because I didn't see dexterous hands.
I saw a robot clumsily picking up hard rigid objects, but the moment it had to do something as delicate as carry a box they gave it gloves so it had enough friction to hold the box without completely crushing it.
The problem is that once you're dealing with breakable objects fingers need a lot of very precise feedback and pressure control to pick things up without damaging them. That's why those dexterous hands, much like his vision only car s
Re: Does he not have youtube? (Score:2)
Boston Dynamics doesn't do sex though. Different market segment.
Re: (Score:3)
Re:Does he not have youtube? (Score:5, Insightful)
If you watch Boston Dynamics robots, it's clear that they are actively balancing themselves all the time, like animals and humans do. If you shove one, it's already running the control loop for balance and it can compensate.
Optimus is much more like older robots that walk using pre-programmed moves. A give-away is how they all kind of crouch a bit, like they are touching cloth, because they need to keep the centre of gravity lower than that of an actual human. When humans walk we are basically falling and catching ourselves, like Boston Dynamics robots do, but that's not what was happening on stage today.
Tesla doesn't have a good track record of hitting Musk's price promises, e.g. with the Model 3.
Also, calling it Optimus wasn't a great decision. Aside from the potential copyright infringement, it makes searching for it conflict with a popular character from a popular franchise.
Re: (Score:2)
Better order some replacement windows so you have them ready when it arrives.
Weaker than his previous demo (Score:2)
In August 2021, he had a Tesla Bot demo. It was faster, and able to dance. Granted, it was a human in a body suit [youtube.com], but it's a rather disappointing reveal when the prototype is a bit conservative in movement and slightly slow.
Compared to Boston Dynamics, which is already having robots moving faster (perhaps a slight jog), able to do minor acrobatics such as a backflip, or even dance.
Re: (Score:1)
Compared to Boston Dynamics, which is already having robots moving faster (perhaps a slight jog), able to do minor acrobatics such as a backflip, or even dance.
A good excuse to trot this video out from 21 months ago: https://youtu.be/fn3KWM1kuAw [youtu.be]
What about the acoustics? (Score:2)
useless demo (Score:1)
It could be quite scary.... (Score:2)
Chauffeur (Score:2)
it features the same technology used to enable the Full Self-Driving beta in Tesla's cars
So stick one of these in a conventional vehicle that has been tricked out with the sensor array you'd get in a Tesla and have you made your average family car into an autonomous (to the level of Tesla's beta) vehicle. Once you'd taught this robot how to refuel it, use the steering wheel and pedals.
Lame (Score:2)
Is it just me? (Score:3)
Or does that robot look an awful like the Terminator?
Re: (Score:3)
Or does that robot look an awful like the Terminator?
A constipated Terminator maybe
If you cannot drive??? (Score:2)
If you cannot create an AI that can drive down a perfectly level, uniform, marked road then you are a century away from humanoid robotics at least.
So what? (Score:2)
The only problem I see is eliminating low paid worker jobs, and that's unfortunately a problem for the elite at the expense of everyone else.
1990 called - they want their robot back (Score:3)
This is seriously lame. They just bolted together of mostly off the shelf parts and got it to barely walk. Musk said his was first time it ever walked without a harness, and he was worried it would face plant.
In the meantime 30 years ago Honda Asimo was running in circles on stage and walking up and down stairs.
University engineering depts are pushing out much more impressive stuff than this ass-cheek clenching shuffling Tesla bot, such as Oregon State's Cassie that just set the world 100m record for a bipedal reobot.
https://interestingengineering... [interestin...eering.com]
Then of course you're got real state of the art in Boston Dynamic's Atlas which has dynamic balance and can do parkour and high kicks when dancing.
https://www.youtube.com/watch?... [youtube.com]
https://www.youtube.com/watch?... [youtube.com]
Seems TeslaBot is about as far from helping in the factory, or being your CatGirl robo-girlfriend (Musk's suggestion) as Tesla is from FSD that doesn't drive into stationary objects.
pitiful (Score:2)
What a pitiful piece of crap. It's a miracle the contraption didn't fall on its face in the demo. Why Musk would want to showcase such junk is beyond me. This is something you could have seen in Japan in 2001, and would have been rightly mocked for being anthropomorphic but useless.
Do you want Kaylon murderbots? (Score:2)
Fuck Elon and his lying bullshit (Score:2)
Oh look, Elon has another new stupid distraction which will never be released, to distract everyone from that he's ten years into selling features for the cost of entire cars, when the features also don't exist
Let's name one of those robots "full self driving"
Stop reporting on this asshole's lies. The Hyperloop was just to kill mass transit. This isn't news.
Sexbot (Score:3)
Brings a whole new meaning to firmware updates...
They've achieved what Boston Dynamics achieved... (Score:2)
barely even twenty years ago.
Why humanoid? (Score:2)
If you want to make a robot to pick strawberries, for example, wouldn't it be better to make it highly optimized for picking strawberries? The human bipedal form is very inefficient and unstable - really the worst choice for something that actually needs to do work.