Blog Tesla Bot Prototype Coming in 2022

Tesla announced Thursday at its “AI Day” that it will build a “Tesla Bot” as soon as 2022.

Chief Executive Elon Musk said the human-like robot will be able to complete tasks that are “boring, dangerous, repetitive jobs that people don’t want to do.”

A Tesla Bot was on display at the event, but as a joke the company put a person in a suit designed to look like Tesla Bot to dance on stage. “Tesla Bot will be real,” Musk said.

Musk said the top goal of Tesla Bot was to make it safe.

“It’s intended to be a friend, of course,” Musk said.

The robot will be 5 feet 8 inches tall and 125 pounds. The head of the robot will include a screen to display useful information. Tesla says the robot will have “human-level hands.” The robot’s top speed will be about 5 MPH.

“At a physical level you can run away from it and most likely overpower it,” Musk said.

The robot leverages the technology developed for its Autopilot feature in its vehicles. The system includes eight cameras and the same computer found in Tesla cars.

The company did not offer any details on pricing for the robot.

 

[JulesVerne]

Why would Tesla suddenly dive into consumer, humanoid bots? At least I assume it's aimed at the consumer market, since it's "intended to be a friend."

I imagine that this will follow the same path as FSD:

1. Make grandiose claims: robot will be able to unload and put away groceries from your car, as well as play grandmaster-level chess. And babysit.
2. Take deposits for robot.
3. Several years late, initial robots are delivered. The "human level" hands have a thumb and two fingers; Elon explains that "hands are harder than we thought. Just look at how hard they are to draw!" They can't walk up or down stairs, but do routinely walk into walls. That's OK, though, because they receive software updates automatically, over the air. First update improves the robot's capability of picking up an egg, which it can now do without breaking the egg more than 80% of the time.
4. Updates add bizarre features that nobody wanted, but get a lot of YouTube time: robot makes loud farting noise when sitting down. "Chill mode" makes robot glance at you, then ignore you, when you give it a command. Robot can play Netflix and Hulu content on its facial screen.
5. A special, limited cadre of influencers gets the "robot can babysit" update. Putting the robot in babysitting mode displays a warning that you're responsible for the baby AT ALL TIMES. Nonetheless many users simply leave the baby alone with the robot, and record what happens with a nanny cam. The resulting YouTube videos range from hilarious to horrifying. There are fatalities.
6. Elon promises general availability of babysitting functions "In a couple of weeks". Months later, Elon says that this will actually come in new, completely re-written from the group up babysitting code V10. Or maybe V11.
7. And that's as far as I can take this, feature parity with FSD!

just to say..... your comment is the best. Thanks for the laugh.

 

[mark95476]

I'm not surprised that we see all of this naysaying on TMC forums. It's been like that since Tesla started so nothing has changed, except that Tesla has been proving the naysayers wrong all that time.

It's been decades since I've looked at robotic dynamics and mechatronics, but they had a lot of really fancy canned algorithms and such for estimating and controlling movement even in a single plane. Processor power was rather limited and you had to be relatively smart to work in mechatronics as it involved lots of complex math and algorithms. The robotics industry was using cameras and other sensors decades before Tesla even existed.

Look at AP in a Tesla:
- Lots of sensors like cameras to identify objects, determine locations, velocity, paths
- Dynamic environment
- Moving through dynamic environment, primarily trying to avoid collisions
- Precise motor control with tight closed/feedback loop to sensors
- Power-limited environment
- Self (battery) powered

All very similar concepts that can be applied to a Tesla Bot and vice-versa to EVs, factories, rockets, and space-ships. I think it was a recruiting event and Tesla probably specifically wants people that work on robotics and mechatronics. Tesla is building things and not just selling search results to advertisers for views.

Sign me up for a Tesla Bot!

I'm surprised at all the naysaying in this thread, and the posters who think it's a joke or a distraction or that Elon wasn't serious. When has Elon ever announced something he wanted to do and then not followed through with the ardent pursuit of said objective? What is Elon's track record with achieving goals so far? Sure he is often behind his own optimistic schedules but that's to be expected when doing things no one else has ever done, isn't it?

I for one think the Teslabot is coming, it will work like Elon says it will, and the entire world economy will be forever changed once it does. Maybe I'm an optimist too like Elon but to me it seems like a matter of when, not if, now?

 

[jsmay311]

People who actually work in robotics are falling over laughing about this. But the Elon fanboys just know in their guts that he’ll make it a success. Not sure who to believe. /s

More importantly, how is no one talking about the dancing man in the spandex robot suit?? Holy ****ing ****! It’s so embarrassing!

 

[cwerdna]

I'm surprised at all the naysaying in this thread, and the posters who think it's a joke or a distraction or that Elon wasn't serious. When has Elon ever announced something he wanted to do and then not followed through with the ardent pursuit of said objective? What is Elon's track record with achieving goals so far? Sure he is often behind his own optimistic schedules but that's to be expected when doing things no one else has ever done, isn't it?

See post #16 in this thread: Blog - Tesla Bot Prototype Coming in 2022.

There are MANY MANY more things at Elon Musk Today and elsewhere that he hasn't followed through on. And, there are many things he's reneged on too.

Tesla mistake? - "fully loaded, newly built cars to our service loaner fleet" (originally posted at Creating the World’s Best Service and Warranty Program | Blog | Tesla Motors) he did do. But that generosity on loaners went away years ago. Now, it's totally YMMV whether you might get any loaner. If you do, it might be a random ICEV. A co-worker who formerly owned a 3 ran into that.

He launched Introducing Tesla Insurance and not surprisingly, it was half-assed with lots of people running into Have Tesla insurance, just got DMV letter "Intent to Suspend" for lack of insurance.

 

[Mengy]

See post #16 in this thread: Blog - Tesla Bot Prototype Coming in 2022.

There are MANY MANY more things at Elon Musk Today and elsewhere that he hasn't followed through on.

But missed dates are not the same thing as "not following through on". Level 5 FSD is in active development, it's happening, that's most of what AI Day was all about.

I feel like some people get hurt because Elon gives dates which are overly optimistic and misses them, but man this stuff isn't EASY you know? And much of it has never been done before by anyone either, right? Just because he is late on robotaxi's does NOT mean they aren't coming. They are.

 

[cwerdna]

But missed dates are not the same thing as "not following through on". Level 5 FSD is in active development, it's happening, that's most of what AI Day was all about.

I feel like some people get hurt because Elon gives dates which are overly optimistic and misses them, but man this stuff isn't EASY you know? And much of it has never been done before by anyone either, right? Just because he is late on robotaxi's does NOT mean they aren't coming. They are.

Nobody has ever said level 5 autonomy was easy, but you followed Elon's grandiose claims over and over and see what others (e.g. Waymo, Cruise Automation + a bunch of Chinese companies) are doing (and their progress), you'd be pretty skeptical of Elon's claims. Not sure what he's really thinking and his intent.

Maybe you're new to all of this, judging by your join date. Seriously, if you've followed this long enough and hadn't drunk the Elon kool-ald, you'll know where I'm coming from.

I've posted about some of this before (e.g. Autonomous Vehicle Progress and Autonomous Vehicle Progress, too bad many links there are dead so you'll need to use archive.org to look at old snapshots).

Elon Musk: EAP solved, on track for FSD completion in 2019 (No one else is close!) has more details on some of the stuff he said on autonomy in the past. Look at where Tesla is now vs. those claims.

 

[Mengy]

Nobody has ever said level 5 autonomy was easy, but you followed Elon's grandiose claims over and over and see what others (e.g. Waymo, Cruise Automation + a bunch of Chinese companies) are doing (and their progress), you'd be pretty skeptical of Elon's claims. Not sure what he's really thinking and his intent.

Maybe you're new to all of this, judging by your join date. Seriously, if you've followed this long enough and hadn't drunk the Elon kool-ald, you'll know where I'm coming from.

You don't know anything about me nor how long I've been reading these forums. You seem to assume a lot about other people, are your feelings about Elon based on emotional assumptions as well?

If you don't know what Tesla's intent is regarding FSD then maybe you need to do more research? Do you believe Tesla will fail at solving level 5 autonomy? And if so, why?

 

[cwerdna]

If you don't know what Tesla's intent is regarding FSD then maybe you need to do more research?

I'm talking about Elon's intent when he throws out his claims, dates, etc. and then misses them but then throws out more that will almost certainly be missed. It's almost like the Charlie Brown and Lucy football gag.

Is he that terrible at estimating? Is he that terrible at not realizing the difficulties and all the edge cases? Is he just trying to troll people? Is he trying to just fool people who don't know any better? How much of it is trying to motivate his team to hit a date by making that date a forcing function? Is it part of some move to boost or prop up TSLA stock price? Is it intended to trick/mislead people into paying for autopilot, FSD, etc.? I can think of dozens of other reasons.

You should look back at the claims that Tesla used on their on web pages when ordering "FSD" back in 2016: All Tesla Vehicles In Production Are Now Equipped With Full Self-Driving Hardware, Model 3 Will Be Too like enhanced-autopilot-self-driving-tesla-autopilot-cost | InsideEVs Photos. They put out publicity stunt videos like at All Tesla Cars Being Produced Now Have Full Self-Driving Hardware which was seriously misleading (CA DMV Report Sheds New Light On Misleading Tesla Autonomous Drive Video). Look at Autonomous Vehicle Disengagement Reports 2016, namely at what Tesla submitted vs. others.

Look at how many other non-autonomy claims related to Tesla he's made and how they never materialized. I already gave you a good source of those.

Again, Tesla and Elon are about hype and their communications style has been "say something and let people's imaginations run wild". Unfortunately, the imaginations often run wild with people who have no technical nor engineering background.

Over the years, I've seen the pattern here and elsewhere of Tesla fanboys who keep believing that Tesla is ahead of everyone else in the autonomy space and/or if not, they will somehow pull ahead of everyone else due to the sheer amount of data they have/can collect. They usually seem to be non-technical people. On the 1st point, nope. On the latter, seems doubtful.

Forget the emotions, the facts speak for themselves.

 

[Soda Popinski]

This is ridiculous. AI is half the battle, just ask Boston Dynamics. They've been working on their robots for decades and they're still nowhere near mass production for at-home use. Musk may be overreaching here.

Looking at Boston Dynamics, I think their problem is they're essentially an R&D company. They're not set up for mass manufacturing, nor do they seem all that interested in it. You can already purchase Chinese made versions of "Spot" for far less than they're selling them for.

 

[Soda Popinski]

People who actually work in robotics are falling over laughing about this. But the Elon fanboys just know in their guts that he’ll make it a success. Not sure who to believe. /s

More importantly, how is no one talking about the dancing man in the spandex robot suit?? Holy ****ing ****! It’s so embarrassing!

I've seen Honda's Asimo, and Atlas (along with other bi-pedal robots) as the DARPA Robotics Challenge a few years ago. I also saw very human hand like prosthetics while there. I don't see why building a humanoid robot would be impossible. My Neato robotic vacuum has a rudimentary LIDAR and can navigate around my house without a problem. Have the vision system powered by a well trained NN for household items should allow a robot to recognize and pick up a grocery bag or maybe even a beer. Voice activated commands are already a thing.

I'm sure it'll be quite terrible, but I don't see why it's impossible. Note the auto OEMs also laughed at Tesla trying to make a mass market EV. Rocket companies also laughed at SpaceX trying to land and re-use a rocket. I don't believe it'll be 2022 though.

 

[FloridaJohn]

I've seen Honda's Asimo, and Atlas (along with other bi-pedal robots) as the DARPA Robotics Challenge a few years ago. I also saw very human hand like prosthetics while there. I don't see why building a humanoid robot would be impossible. My Neato robotic vacuum has a rudimentary LIDAR and can navigate around my house without a problem. Have the vision system powered by a well trained NN for household items should allow a robot to recognize and pick up a grocery bag or maybe even a beer. Voice activated commands are already a thing.

I'm sure it'll be quite terrible, but I don't see why it's impossible. Note the auto OEMs also laughed at Tesla trying to make a mass market EV. Rocket companies also laughed at SpaceX trying to land and re-use a rocket. I don't believe it'll be 2022 though.

To me, the biggest problem is power. How much space and weight is required to power up the bot for a reasonable amount of time? All the space taken up with batteries, etc is less room for mechanisms, sensors and electronics. Most of the bipedal robots I’ve seen are wearing some type of backpack-style attachment to solve this problem.

 

[Soda Popinski]

Most of the bipedal robots I’ve seen are wearing some type of backpack-style attachment to solve this problem.

If we're talking about Atlas, we should remember Atlas is hydraulically powered. That means a hydrolic pump, and a motor big enough to power it along with the batteries. Atlas is also incredibly strong. Strong enough to do somersaults. As long as TeslaBot doesn't need to do parkour, I think it can get away with lower power consumption and lower performance actuators instead of hydraulics.

If we look at ASIMO, its got a runtime of an hour. For a household robot that can plug itself in (like my vacuum), it can get away with an hour of use at a time. If it's not actively moving or moving a lot, that time can stretch out further.

 

[linux-works]

to save power, it will need to have multiple levels of sleep states.

question is: how grouchy will it be once its woken from one of those states, against its will?

we talking cat grumpy or even more?

 

[thohell]

re-use of the 'AI tech'

but really, that's not even close to a ready to divert attention and funds away from the car business.

are all the bugs fixed? are enough service centers built in all states? charging stations for the new non-T's that will come in? how about making parts so that people dont wait weeks for a repair? maybe build extra cars so that you can give loaners?

no. none of that.

lets build stupid robots that are just marketing.

how many people are on this? how much funds that could be used in real world, today-problems, with the cars?

he constantly bites off more than he can chew.

I just wish he'd put all that energy toward THIS field. cars. and get the body panels right, the build quality right, all the boring things that other vendors figured out. too boring for you, elon?

time for him to retire and let someone focused, take over.

Tesla... isn't...a...car...company.... Tesla would die without Elon.

 

[Soda Popinski]

Tesla... isn't...a...car...company.... Tesla would die without Elon.

Why do you think they're making the Teslabot? Coupled with Neuralink, he's going to upload his consciousness into Dojo with the robots as his bodies, like Ultron.

 

[Dan D.]

Why do you think they're making the Teslabot? Coupled with Neuralink, he's going to upload his consciousness into Dojo with the robots as his bodies, like Ultron.

Oh good, then maybe a little bit of Elon's consciousness in every FSD. "Hey car, when am I going to arrive at Walmart?"

"Two weeks!"

 

[SarahsDad]

If Teslabot ever does happen, you know Elon will make "Emissions Testing Mode" an option (starts at 5:10)

 

[Meteile]

Looking at Boston Dynamics, I think their problem is they're essentially an R&D company. They're not set up for mass manufacturing, nor do they seem all that interested in it. You can already purchase Chinese made versions of "Spot" for far less than they're selling them for.

Right but how useful is spot though? To my knowledge there are no humanoid androids for sale similar to what BD is offering.

 

[Akikiki]

Right but how useful is spot though? ...

Stars in the War of the Worlds series on Epix

 

[Meteile]

Stars in the War of the Worlds series on Epix

And a certain Black Mirror episode