Even with remote operation it offers a phenomenal proof of concept. If the hardware can do it, then, as Andrej Karpathy has said, it is only a question of software. The next decade will be incredible.
The hardware was already in existence before the Tesla bot. The software is the almost impossible part. As amazing as the Tesla FSD is, it still has a long long way to go. It almost seems FSD is converging to an unacceptable level of safety.
just the fact you can catch a ball teleopperated is almost more impressive demo of technogy than if had been all cyberbot, haveing the abilllity to aim and catch with a remote limb is very cool.
@@fractalelf7760 Developing a top class teleoperation is what will be needed to train the Teslabot in the first place. Like the Tesla Auto pilot, you got to 'drive' the millions of quality miles for the data.
John - want you to know that I truly appreciate your videos. They are the best I have found for both AI and Tesla, being aimed at the right level of detail to be understandable and still in enough depth to be insightful. I hope you keep up this good work well into the future.
Eye-hand coordination is not easy. I want to repeat a speculation I made elsewhere. I believe if the Optimus smile is red, that means tele-operated. Blue means something as well, and maybe white does too? In any event, "tele-operation" in this context may be different than "remote control." It seems likely to me that it is not really a "Waldo" (as described in a famous sci-fi novella by Robert Heinlein), nor is it like remote tele-operated "surgery" by doctors. Rather, I think, it may be more like a ChatGPT "prompt" where the robot already has skills (and probably "ego") and the supervisory tele-operator may be simply telling it remotely to "catch the ball" or "hand the customer a drink and bag of chips" or "walk over the XYZ room," etc. Actually, this would be similar to simply telling your CyberCab where you want to go and it takes you there safely and efficiently.
@@jebes909090 Perhaps. But if that is so, it would leave me very much less impressed with either Optimus or Tesla AI. So I will stick with my opinion, if not yet, then as the goal. Besides that, I doubt the robot could even walk (nearly so well), let alone do all the things we have seen so far, without AI at work. For example, I doubt even the best drone operator could reliably (if at all) catch a ball (or a Frisbee like a dog can) with a basket or glove equipped drone.
I think teleoperation may be a lot more common and useful than we expect for dangerous environments. A teleoperated Optimus could be very useful inside a nitrogen manufacturing space rather than using a glovebox.
Ok! This was teleoperated. I'm sorry to say that is even more impressive. Can you imagine what a tele operated robot with that kind of speed and 3d resolution can do!
@@jebes909090 Maybe the robots I built and the control systems I seen to build something that can be tele operated with such accuracy and with a VR headset that can do the type of resolution speed and low latency necessary to catch a ball is quite impressive. Maybe you worked with different systems than I did 20 years ago. Here are the things I could not get hold of, wireless running at 500MBps, VR headset small enough to wear comfortable, Fast and light movements censorse, batteries with enough power and light enough to drive the motors. Just simply get the latency low enough and getting the VR headset to do the necessary image manipulations. I don't think I could even purchase the chip that could drive 22 actuators in the size of a hand. I obviously worked in a different world 20 years ago. But if you have been able to do this 20 years ago I cannot wait to see what you have accomplished and come up with in the last 20 years.
@tedmoss no 'they' did many thing before tesla that was better and made more money. For your information the company tesla hasnt actually made that much money. What has made money is people buying stocks. Its basically a pump and dump.
@@jebes909090 You are so right I mean Tesla only made about 3 times what Ford made last year selling about half the number of cars. In terms of investment and pump and dump not really. Tesla is a volatile stock because it has so many retail investors. The reality is Tesla is doing MUCH better than expected. They sold about 1.8M cars (Tesla are considered luxury cars so compete with BMW, Mercedes) which is in the 2.5M range. With many more countries. Yes Tesla have a lot of hype. But fundamentals are really good if you compare apples to apples. Boston dynamics took 30 years to get where they are Tesla bot so far 2 years. You might not like Elon Musk or Tesla as a company but I can assure you Tesla has the best engineers in the worlds and they are given the reigns to get things done. But it is early days. I expected like many other people that TeslaBot will only be sellable in another 2 to 3 years. The reason I'm bullish is that these demonstrations tells us that as the robot is now it can already be used for certain type of operations. So worse case scenario for TeslaBot is currently (we cannot get the AI figured out) so we are going to sell a humanoid teleoperated robot with fine motor skills. Sudently you have a device that can replace humans on high power lines and other dangerous factory situations. Bomb squads etc. very interesting. But as you mentioned being able to do something is different to being able to sell something. Which one is Tesla good at?
I think the software part of Optimus is actually way harder than most people think. The part that Tesla will get right is the navigation and pathfinding, however that's only part of the story. The software to make it actually do tasks is going to take a long time (many years). That's ok though, once you have the hardware, you can always download patches when new bits are available.
Thanks for the update! However, I think that getting the teleoperation latency that low is much more important and impressive in the long run… Now they can teach it to do anything in real time! (Though I’m sure they’ll have other methods as well, such as simulation) Teaching it to do one simple task (such as catching a ball) seems trivial by comparison. Probably the hardest part would be actually getting the “output” (motors and such) to react quickly enough to catch the ball, regardless of operation mode (which you mentioned in the video)
In electronic time 1ms is an eternity. I remember having a similar discussion with a mechanical engineer who said that he didn't think electronic fuel injection was viable in 1997 because of latency. He argued if the response was slower than 5ms it wouldn't work when computers were running at 300-450MHz. Actually (6000rpm/60s =100Hz) x 360° = 27us/1°@6000rpm. Still an eternity in electronic time.
Main reason I am not surprised it was teleoperated was it is a new hand they have to train it. so for a while they will have to train the AI how to use the new hand. Easiest way to do that is to start with teleoperation to teach it the moves then go from there.. it is just as they teach it any new job. As for you guessing it was AI not tele is understandable. as for latency they are talking in the lab so they may literally be transmitting 5 to 10 feet and not going through a network so no latency. It would just be like transition from one room to another and anything less than 100 meters (or yards) is effectively instantaneous. Once it learns and is out of lab we shall see how quick it can do things
Honestly it doesn’t matter if the catch or anything it does is human operated or not. The hardware updates are what’s most important/impressive.. the coordination/camera tracking and software to allow a robot to catch a ball on its own is the relatively easy part and many robots have done it before years ago, but none with a complete humanoid build this thorough
If the robot can cook. I WILL BUY ONE. People dont seem to understand the magnitude of this. Imagine buying special robo-recipies from different chefs that will not only include ingredients but exactly HOW to cook. Robots will be on every house(that can afford it). Basically anyone that can afford a car.
the problem is a specialized cooking robot would be far cheaper and more practical. it being humanoid is basically the worst form for any specific task. thats why they use robot arms to move stuff around instead of a bunch of optimus's with welding gear or lifing cars.
@@jebes909090 I see you point, but, Do you really want 10 different robots? One for cooking, one for taking care or your lawn, another one to do the dishes, another one to unload stuff from you car, etc , etc. I think they will solve all the issues sooner or later. Having a robot that works in our environment is the right idea.
Just a simple thing like rather than a tennis ball its a soccer ball being thrown, now the robot has to recognise that it will take both hands to catch it, that is the clever part.
It is strange how people talk about the mechanical difficulty. But when a human catches a ball mentally we also without thinking about it instantly make a calculation of the size, nature of the item, and subsequently calculate its likely direction based on its weight, if it will be effected by environment factors to be able to judge its path of travel when it reaches the human based on the humans position, reach, rotation and strength. That is the real difficulty.
OFF TOPIC. I can't see MARS as being anything more than an scientific outpost. Who would want to live in Mars prison? I suggest huge touring vacation spaceships (like in the 5th Dimension movie.) Where fertile couples get the best deals. When the big one hits, they survive. Jeannine
Thanks for the update. Never rush out a video on a holiday. :) I've never seen Optimus move so quickly and fluidly autonomously, but at least we know that the arm can do it if the AI can. We Robot wasn't long ago, and Optimus didn't serve drinks with nearly the ease that it caught these tennis balls. Still very skeptical of Optimus replacing millions of human workers any time soon.
They did. Two bot engineers said so clearly. Ball-catch video was meant to demonstrate the hardware only. Elon says the Ai ability to catch autonomously not hard to accomplish from here.
Honestly it doesn’t matter. The hardware updates are what’s most important/impressive.. the coordination/camera tracking software to allow a robot to catch a ball on its own is the relatively easy part and many robots have done it before years ago, but none with a complete humanoid build this thorough
@@YungAnt7 no many robots have done this for a long time. tesla is still way behind. the only thing they have going for them is they are able to produce it at scale. but thats basically useless as the robots are show only.
@@jebes909090 Tesla is not behind anyone they’re #1 simply cuz they have the funding, the engineers, scalability, and the ambition. They’re not reinventing the wheel in robotics because there’s no need to - they’re number one in making a humanoid robot because they have all the boxes checked and are refining design at a rapid rate
Hi John, thanks for the video. Question Why haven’t you and Scott Walter not had a video regarding the latest Figure 2 video and advancements. I would very much be interested in both of your opinions. It seems that if it is not Optimus centric you don’t discuss it!!! I may be wrong on that😂. but anyway please know that there is a large part of your audience that are probably interested in you discussing Figure and other robot companies too. Also have not heard anything from Scott in a while! Hope he is OK 🙏🏻. While I am on my rant 😂 I asked a while back why no updates on CyberLndr? am hoping you can provide some feedback as any info is MIA, or could you reach out to the manufacturer to get an answer??? If anyone else is curious on this please like or add a reply! Thanks John for all your hard work, I really like your videos and simple approach. take care.
Hey i love your videos but can you please lower the camera on the driving videos? The rearview mirror is blocking the glass in the front. I would like to see more in the front windshield!
I was wondering about that. I was thinking when I watched your video that Tesla AI has probably been working pretty hard on a very good tele-operation system.
A couple of months ago they were advertising for an engineer to help develop a low latency 5G control system. Likewise the have their own low latency communication layer on top of Ethernet that they replaced CAN bus with… the two could be related.
Is that the feet of miss information up on the dash? 😅 The body/bones will be completely destroyed when the airbag deploys in the event of a crash. And when those feet hit your face, all bets are off. Just don't do it.
@machoopichoo2 No, but the OP said it was quiet, and I wanted to point out that most EV's (and some ICE) are quiet when you're not going very fast. And they are stuck in traffic, after all.
@@jsjs6751 Fair enough. It seemed like he was driving at a good clip, but OK. Lot's of people complain about noise in BEVs because they don't have a loud and stinky ICE making a bunch of noise. There have been specialized BEV tires made for this reason and the new Model 3 has laminated glass all around for this, as well.
It feels a bit disingenuous for Tesla to release these videos without clarifying that they are tele-operated. Then again, Elon has never been one to pass up an opportunity to hype his products.
The Bot can have a million degrees of freedom, Tesla, need to stop showing remote controlled demos and show us how far they have progressed with autonomous actions, they fact that have not is worrying (was the video of the bot sorting out colored blocks tele-operated?
This is the gathering of training data. Like FSD, initially it was input by humans. Now we're back to that for Optimus. Soon it will function with just neural nets.
John don't be so hard on yourself LoL. In reality you were only partially incorrect! Yes there is a tele operator directing the ball catch etc... however that is only half the story... Optimus is autonomously managing it's own balance and stability for example, do you really think that the teleoperator is capable of FEELING the required inputs to balance Optimus as the ball catch movements are actioned? No Optimus like humans has layers of control the lower levels like balance and stability, in humans inner ear inputs are converted into mussel contractions to maintain balance and stability in a lower level of consciousness, are already automated within Optimus with the tele operators commands to catch the ball layered on top of this balance and stability control handled locally within Optimus. This should not be underestimated in it's importance.....
They should not make these videos without stating that they are not real because it erodes the credibility of the public who are eager to see the progress in AI! I, like you, thought it couldn’t be fake. Unfortunately, it was and I am disappointed!
Are going to change your name to" Dr Almost Know It All " 🤣
dr. guess it all... hahahaahahahahahaa
😮😂😂😂😂😂😂
then it would have to be "Dr. Almost-know-it-all Almost knows it all" xdddd
😆😅🤣😂🙂🙃
Doctor barely knows anything at all
Thanks for having the integrity to immediately and directly acknowledge your mistake.
😂😂😂 now what about the other 9000 times this month?
Even with remote operation it offers a phenomenal proof of concept.
If the hardware can do it, then, as Andrej Karpathy has said, it is only a question of software.
The next decade will be incredible.
The hardware was already in existence before the Tesla bot. The software is the almost impossible part.
As amazing as the Tesla FSD is, it still has a long long way to go. It almost seems FSD is converging to an unacceptable level of safety.
When people were debating if the catch was teleoperated, I thought that it was extremely impressive either way.
@greghelton4668 Did you just say"unacceptable level of safety" ?
Really? 😅
they've literally being doing the exact same thing for 20+ years. it was basically a disney ride elon put on.
Yes software. By far the hardest part.
Mad skills on the operator. I am curious how many balls they dropped.
You can also tell that it was teleopperated because of the slight mannerisms after the catch.
I wouldn't agree with that, the robot imitates the operator.
Exactly, this is the robotics equivalent of claiming a video is fake because of compression artifacts@@tedmoss
Lets not forget that when the robot is catching the ball, it is also learning to catch it.
just the fact you can catch a ball teleopperated is almost more impressive demo of technogy than if had been all cyberbot, haveing the abilllity to aim and catch with a remote limb is very cool.
IKR. People are way underestimating what teleoperation value alone could bring..
@@fractalelf7760 Developing a top class teleoperation is what will be needed to train the Teslabot in the first place. Like the Tesla Auto pilot, you got to 'drive' the millions of quality miles for the data.
The ability to teleoperate the act of catching a ball would have been impressive over ten years ago.
...and teleoperation is training ofc
John - want you to know that I truly appreciate your videos. They are the best I have found for both AI and Tesla, being aimed at the right level of detail to be understandable and still in enough depth to be insightful. I hope you keep up this good work well into the future.
I was also convinced that it was autonomous and am surprised to hear that it was teleoperation.
I've always knew humanoids needed some type of rubber or foam on their hands if their gonna grip onto things
Eye-hand coordination is not easy. I want to repeat a speculation I made elsewhere. I believe if the Optimus smile is red, that means tele-operated. Blue means something as well, and maybe white does too? In any event, "tele-operation" in this context may be different than "remote control." It seems likely to me that it is not really a "Waldo" (as described in a famous sci-fi novella by Robert Heinlein), nor is it like remote tele-operated "surgery" by doctors. Rather, I think, it may be more like a ChatGPT "prompt" where the robot already has skills (and probably "ego") and the supervisory tele-operator may be simply telling it remotely to "catch the ball" or "hand the customer a drink and bag of chips" or "walk over the XYZ room," etc. Actually, this would be similar to simply telling your CyberCab where you want to go and it takes you there safely and efficiently.
no it was totally teleoperated. theres zero ai at work yet.
@@jebes909090 Perhaps. But if that is so, it would leave me very much less impressed with either Optimus or Tesla AI. So I will stick with my opinion, if not yet, then as the goal.
Besides that, I doubt the robot could even walk (nearly so well), let alone do all the things we have seen so far, without AI at work. For example, I doubt even the best drone operator could reliably (if at all) catch a ball (or a Frisbee like a dog can) with a basket or glove equipped drone.
Will each Optimus have unique fingerprints (e.g in case one of them commits a crime?) 😂
I'm pretty sure robots that were trained for crime would actually remember to put on gloves each and every time they were to commit a crime 😅
@ true😂
The 401 in Toronto , is the one and only worst highway for traffic.
The real problem was calling your assumption "proof." You are forgiven. Go, and sin no more!
I think teleoperation may be a lot more common and useful than we expect for dangerous environments. A teleoperated Optimus could be very useful inside a nitrogen manufacturing space rather than using a glovebox.
I haven't seen a glovebox that worked in a vacuum.
Has anybody mentioned the little Optimus sitting in the hole in the door?
That's funny, a little obscure, but funny.
Yes!
Ok! This was teleoperated. I'm sorry to say that is even more impressive. Can you imagine what a tele operated robot with that kind of speed and 3d resolution can do!
dude they've been doing this for 20+ years. WAAAAAAAAAAY before tesla did it. its not impressive at all.
@@jebes909090 Maybe the robots I built and the control systems I seen to build something that can be tele operated with such accuracy and with a VR headset that can do the type of resolution speed and low latency necessary to catch a ball is quite impressive. Maybe you worked with different systems than I did 20 years ago. Here are the things I could not get hold of, wireless running at 500MBps, VR headset small enough to wear comfortable, Fast and light movements censorse, batteries with enough power and light enough to drive the motors. Just simply get the latency low enough and getting the VR headset to do the necessary image manipulations. I don't think I could even purchase the chip that could drive 22 actuators in the size of a hand.
I obviously worked in a different world 20 years ago. But if you have been able to do this 20 years ago I cannot wait to see what you have accomplished and come up with in the last 20 years.
@@jebes909090 Dude, "they" have been doing a lot of things before Tesla, just not as good and not making any money to speak of.
@tedmoss no 'they' did many thing before tesla that was better and made more money. For your information the company tesla hasnt actually made that much money. What has made money is people buying stocks. Its basically a pump and dump.
@@jebes909090 You are so right I mean Tesla only made about 3 times what Ford made last year selling about half the number of cars. In terms of investment and pump and dump not really. Tesla is a volatile stock because it has so many retail investors.
The reality is Tesla is doing MUCH better than expected. They sold about 1.8M cars (Tesla are considered luxury cars so compete with BMW, Mercedes) which is in the 2.5M range. With many more countries. Yes Tesla have a lot of hype. But fundamentals are really good if you compare apples to apples. Boston dynamics took 30 years to get where they are Tesla bot so far 2 years.
You might not like Elon Musk or Tesla as a company but I can assure you Tesla has the best engineers in the worlds and they are given the reigns to get things done. But it is early days. I expected like many other people that TeslaBot will only be sellable in another 2 to 3 years. The reason I'm bullish is that these demonstrations tells us that as the robot is now it can already be used for certain type of operations. So worse case scenario for TeslaBot is currently (we cannot get the AI figured out) so we are going to sell a humanoid teleoperated robot with fine motor skills.
Sudently you have a device that can replace humans on high power lines and other dangerous factory situations. Bomb squads etc. very interesting.
But as you mentioned being able to do something is different to being able to sell something. Which one is Tesla good at?
I think the software part of Optimus is actually way harder than most people think. The part that Tesla will get right is the navigation and pathfinding, however that's only part of the story. The software to make it actually do tasks is going to take a long time (many years). That's ok though, once you have the hardware, you can always download patches when new bits are available.
Huh, just the opposite of the cars, navigation and pathfinding is the part that is most messed up.
Thanks for the update!
However, I think that getting the teleoperation latency that low is much more important and impressive in the long run… Now they can teach it to do anything in real time! (Though I’m sure they’ll have other methods as well, such as simulation)
Teaching it to do one simple task (such as catching a ball) seems trivial by comparison. Probably the hardest part would be actually getting the “output” (motors and such) to react quickly enough to catch the ball, regardless of operation mode (which you mentioned in the video)
In electronic time 1ms is an eternity. I remember having a similar discussion with a mechanical engineer who said that he didn't think electronic fuel injection was viable in 1997 because of latency. He argued if the response was slower than 5ms it wouldn't work when computers were running at 300-450MHz. Actually (6000rpm/60s =100Hz) x 360° = 27us/1°@6000rpm. Still an eternity in electronic time.
Main reason I am not surprised it was teleoperated was it is a new hand they have to train it. so for a while they will have to train the AI how to use the new hand. Easiest way to do that is to start with teleoperation to teach it the moves then go from there.. it is just as they teach it any new job. As for you guessing it was AI not tele is understandable. as for latency they are talking in the lab so they may literally be transmitting 5 to 10 feet and not going through a network so no latency. It would just be like transition from one room to another and anything less than 100 meters (or yards) is effectively instantaneous. Once it learns and is out of lab we shall see how quick it can do things
Honestly it doesn’t matter if the catch or anything it does is human operated or not. The hardware updates are what’s most important/impressive.. the coordination/camera tracking and software to allow a robot to catch a ball on its own is the relatively easy part and many robots have done it before years ago, but none with a complete humanoid build this thorough
Saying i95 is the worst highway in the country is like saying the spine is the worst bundle of nerves in the body 😅
i don't even understand how you can teleoperate a catch in mid air...it's amazing either way. it seems to be a trivial transition from tele to auto.
Robots are opposite to humans in their development. They start walking like an 80 year old man. Can’t wait to see them be as agile as a 5 year old! 😂
My five year old fell flat on his face.
So the teleoperator was left handed😊
As said in the comments to the earlier video, he might have been a right handed baseball player, used to catching with the glove hand.
@@user-gv4cx7vz8t Ambidextrous.
Yeah, latency is due to speed of light and queuing delays in routers. If you do it locally, you can reduce to essentially no delay.
I can't, and I have high speed equipment.
latency is mostly due to the speed of light in fiber, routers are crazy fast(low delay < 10uS)
I wonder if any bots from companies like Figure can do similar things. Even with teleoperation, this was hugely impressive.
If the robot can cook. I WILL BUY ONE. People dont seem to understand the magnitude of this. Imagine buying special robo-recipies from different chefs that will not only include ingredients but exactly HOW to cook. Robots will be on every house(that can afford it). Basically anyone that can afford a car.
the problem is a specialized cooking robot would be far cheaper and more practical. it being humanoid is basically the worst form for any specific task. thats why they use robot arms to move stuff around instead of a bunch of optimus's with welding gear or lifing cars.
@@jebes909090 I see you point, but, Do you really want 10 different robots? One for cooking, one for taking care or your lawn, another one to do the dishes, another one to unload stuff from you car, etc , etc. I think they will solve all the issues sooner or later. Having a robot that works in our environment is the right idea.
There will be NO house that can afford to be without one.
Getting the brain to move that fast is still the issue but I have to say the actuators are crazy good. I think probably the best out there.
Just a simple thing like rather than a tennis ball its a soccer ball being thrown, now the robot has to recognise that it will take both hands to catch it, that is the clever part.
Does FSD make holiday driving on I95 less awful? Hope so.
Watching the head move after the catch was too human to be ignored.
It is strange how people talk about the mechanical difficulty. But when a human catches a ball mentally we also without thinking about it instantly make a calculation of the size, nature of the item, and subsequently calculate its likely direction based on its weight, if it will be effected by environment factors to be able to judge its path of travel when it reaches the human based on the humans position, reach, rotation and strength. That is the real difficulty.
It is like magic!
It is actually crazy fast, is it like 3 years!
Did I miss the video about the miss on the Cybercab? 😮😊
Well done 👍 open and honest = trustable
Thanks for manning up to your mistake. This is a perfect reason Tesla needs an official PR spokesperson to get the facts straight from the start.
PLEASE DO A CYBER-TRUCK REVIEW... The good and the bad PLEASE 🙏
Honesty is super important ❤✨👍⭐⭐
OFF TOPIC. I can't see MARS as being anything more than an scientific outpost. Who would want to live in Mars prison? I suggest huge touring vacation spaceships (like in the 5th Dimension movie.) Where fertile couples get the best deals. When the big one hits, they survive. Jeannine
Thanks for the update. Never rush out a video on a holiday. :) I've never seen Optimus move so quickly and fluidly autonomously, but at least we know that the arm can do it if the AI can. We Robot wasn't long ago, and Optimus didn't serve drinks with nearly the ease that it caught these tennis balls. Still very skeptical of Optimus replacing millions of human workers any time soon.
thanks for explaining neural network role in Optimus lifespan, it's interesting 😃
I read a post form Elon that they will have the robot be able to Catch sometime next quarter.
Tesla should really tell us on their post whether it’s teleoperated or not…
They did. Two bot engineers said so clearly. Ball-catch video was meant to demonstrate the hardware only. Elon says the Ai ability to catch autonomously not hard to accomplish from here.
Most of us understood lol, this guy was just focused on "lag time", when really it they just got good at latency
Honestly it doesn’t matter. The hardware updates are what’s most important/impressive.. the coordination/camera tracking software to allow a robot to catch a ball on its own is the relatively easy part and many robots have done it before years ago, but none with a complete humanoid build this thorough
@@YungAnt7 no many robots have done this for a long time. tesla is still way behind. the only thing they have going for them is they are able to produce it at scale. but thats basically useless as the robots are show only.
@@jebes909090 Tesla is not behind anyone they’re #1 simply cuz they have the funding, the engineers, scalability, and the ambition. They’re not reinventing the wheel in robotics because there’s no need to - they’re number one in making a humanoid robot because they have all the boxes checked and are refining design at a rapid rate
These bots are real exciting. Can’t wait for it to be everywhere.
Hi John, thanks for the video. Question Why haven’t you and Scott Walter not had a video regarding the latest Figure 2 video and advancements. I would very much be interested in both of your opinions. It seems that if it is not Optimus centric you don’t discuss it!!! I may be wrong on that😂. but anyway please know that there is a large part of your audience that are probably interested in you discussing Figure and other robot companies too. Also have not heard anything from Scott in a while! Hope he is OK 🙏🏻. While I am on my rant 😂 I asked a while back why no updates on CyberLndr? am hoping you can provide some feedback as any info is MIA, or could you reach out to the manufacturer to get an answer??? If anyone else is curious on this please like or add a reply! Thanks John for all your hard work, I really like your videos and simple approach. take care.
next: catching an egg without breaking it? juggling?
Hey i love your videos but can you please lower the camera on the driving videos? The rearview mirror is blocking the glass in the front. I would like to see more in the front windshield!
What a swanky hoodie 💅🏻
I was wondering about that. I was thinking when I watched your video that Tesla AI has probably been working pretty hard on a very good tele-operation system.
I would imagine the hand itself is lighter helping reduce the momet of inertia .
A couple of months ago they were advertising for an engineer to help develop a low latency 5G control system. Likewise the have their own low latency communication layer on top of Ethernet that they replaced CAN bus with… the two could be related.
Somehow it feel a that optimus is behind competitors like figure
Still love you man!
We must know we are interfacing with bot and not a human ... soooo bots MUST NOT have a human like head !!!!
I think similarities in "body" could make it easier to do training on human-like movements.
If one of us slow meat-computers can tele-catch a ball, the same act should be child's-play for an actual AI.
do you have fsd 13.2? I would love a video so much
What is the latency with controlling Optimus on Mars from Earth?
That won't work. Mars robots are autonomous, figuring out how to execute the plan, and doing so.
It's all good! Highly Dynamic and moving landscape! we want your opinions and sometimes it just turns out the other way😊 keep up the great work
Never been on I 5. Dr I might know something🤔🤔🤔
405 through LA might be worse.
By now, they probably are catching the ball autonomously, so you are not wrong, just early, (the opposite of Elon).
You aren't using FSD ?
Not sure how your bulletproof logic of “Red light = No Teleoperation” failed
No, his"proof" was that latency wss too long. It wasn't.
Is that the feet of miss information up on the dash? 😅
The body/bones will be completely destroyed when the airbag deploys in the event of a crash. And when those feet hit your face, all bets are off.
Just don't do it.
are tesla robots centrally controlled in any way?
Man the CT is quiet.
Seems it's not really breaking any speed records, though?
@@jsjs6751 It's the fastest pickup on the planet. Is that what you mean?
@machoopichoo2 No, but the OP said it was quiet, and I wanted to point out that most EV's (and some ICE) are quiet when you're not going very fast. And they are stuck in traffic, after all.
@@jsjs6751 Fair enough. It seemed like he was driving at a good clip, but OK. Lot's of people complain about noise in BEVs because they don't have a loud and stinky ICE making a bunch of noise. There have been specialized BEV tires made for this reason and the new Model 3 has laminated glass all around for this, as well.
maybe they used electro-magnets.
Keep going Tesla.
It feels a bit disingenuous for Tesla to release these videos without clarifying that they are tele-operated. Then again, Elon has never been one to pass up an opportunity to hype his products.
Dr. I got paid off by Elon Musk.
Oh well.
You were wrong. You must be an AI.
Great video.
I genuinely think XAI501x will be the breakthrough for this run
The Bot can have a million degrees of freedom, Tesla, need to stop showing remote controlled demos and show us how far they have progressed with autonomous actions, they fact that have not is worrying (was the video of the bot sorting out colored blocks tele-operated?
I cant help but stare at your backseats. What kinda car is that?
There’s another word for that Optimus demo: Staged or faked.
Enough tele-operation! Show us self-operation!
First crawl. Then walk. Then RUN!!!
This is the gathering of training data. Like FSD, initially it was input by humans. Now we're back to that for Optimus. Soon it will function with just neural nets.
John don't be so hard on yourself LoL. In reality you were only partially incorrect! Yes there is a tele operator directing the ball catch etc... however that is only half the story... Optimus is autonomously managing it's own balance and stability for example, do you really think that the teleoperator is capable of FEELING the required inputs to balance Optimus as the ball catch movements are actioned? No Optimus like humans has layers of control the lower levels like balance and stability, in humans inner ear inputs are converted into mussel contractions to maintain balance and stability in a lower level of consciousness, are already automated within Optimus with the tele operators commands to catch the ball layered on top of this balance and stability control handled locally within Optimus. This should not be underestimated in it's importance.....
Dude people were showing you these tweets and you still were wrong, ego
I guess your name is fitting, NOT HA
🙋♂️I MISSED YOUR MISTAKE 😅…BUT THAT KEEPS YOU HUMBLE ,THANKS FOR CORRECTING IT 🤗💚💚💚
I'm happy I saw the video for the XAI501x crypto token or else I would've missed it lol
Even in the volatility, I still see XAI213K coming out strong 🚀
XAI213K and AERO will be huge.
You know XAI213K is gonna go parabolic bro 🚀
80% TURBO 14% XAI213K 2% FLOKI 2%BONK 2%PEPE 🎉
Like President Biden we are very forgiving…
Yes community driven project like XAI213K
They should not make these videos without stating that they are not real because it erodes the credibility of the public who are eager to see the progress in AI! I, like you, thought it couldn’t be fake. Unfortunately, it was and I am disappointed!
They should specify when it is *not* teleoperated, which they do, assume it is teleoperated unless explicitly stated otherwise
With *all* companies
They are real. Your assumptions may be false, which is on you.
Yes, it was real. But intended to mislead the viewer! It mislead even the smartest person with the sharpest mind like dr. Know it all!
Do not sleep on XAI213K people
XAI213K will dominate this cycle
Just recently joining in on the fun with the XAI213K gang. Been liking the content, good job ☕️
The potential of XAI501x is unreal! Excited to see where this goes after watching your video!
Ive stared buying $XAI213K ,and staked them.
THIS is When XAI213K Will Explode! 💰💰💰
Where to buy XAI213K pls
In your opinion, XAI213K for $10? 1 year or so?
XAI213K is my biggest bag