Tesla Optimus vs. Boston Dynamics Atlas (Watch the Reveals)

Tesla just revealed a working prototype of its new Optimus robot. See how it compares to the already established Boston Dynamics Atlas robot.

[ezequiasluiz]

Tesla's Optimus, the only robot that can simulate a herniated disc like an old man

[DB-umrb]

There’s an interesting East Coast vs West Coast style buried in this. Boston stuff just shows up on YouTube. California stuff has a whole ceremony like it’s the academy awards or something…

[RicoCilliers]

lol boston dynamics out here doing literly gymnastics with their machines while Tesla is basically glorified animatronics

[Gumston]

Imagine if Elon just gave 40 billion to Boston dynamics instead of buying twitter.

[hwaht]

Whenever I see the Boston Dynamics stuff, it just blows me away. Sometimes, robots CAN be cool!

[OweEyeSea]

The Atlas is definitely farther along than Optimus, but to call these clips "Reveals" is misleading. That parkour clip at the end is a full 8 years after the initial Atlas reveal, and 12 years after the earlier PETMAN project Atlas was based on. Boston Dynamics has been working on robotics for 30 years. So when you consider how far Tesla has gotten this in a year, it's pretty impressive.

[sherrypeterson]

Love what Teslas done in such a short amount of time. The hands, stream lined body, FSD AI and cost of Optimus is exciting. I’m sure that more is yet to come.

[biz]

It's safe to say Boston dynamics wins this round

[JoeyCbr]

@5:20 That robot is seriously going to get revenge

[ewaf]

Well although not being able to walk Optimus can wave his hand.

Incredible

[motherbrain]

I'm a huge fan of Boston Dynamic since their first robot dog, but they're not working on a commercial product. Tesla on the other had is purely working and OPTIMIZING for a consumer product. I wish the Boston Dynamics expertise on locomotion would benefit Tesla's bot. Both expertise together would create something amazing.

[rkgsd]

It's the battle over which company's robot will take your jobs first.

[toastedmatt]

While A.I is gonna be revolutionary for these robots, I think a lot of people are overlooking a very simple and practical application for them, and that is to control them with VR. If we put a little more research into things like VR gloves with haptic feedback and Omni-directional treadmills (which will both be much easier to make then A.I in the short term), we could drop a bunch of these things onto the surface of the moon and have people control them from Earth. If lag is an issue we could even send a team of astronauts to the Lunar Gateway and have them control the robots on the surface, using them to design a fully operational lunar base without needing to put any actual humans on the surface. “Cybernauts” if you will.

[haili]

Honda and Boston Dynamics started building their robots 30+ years ago when there wasn’t AI and FSD. Tesla started just last year but they did it at the right time because it has AI, FSD and the best computing power.

[TheBendixSA]

The walk and balance of the atlas bot as well as deg./sec movement is really looking good. It's also very impressive what Tesla has done in basically no time at all.

[DemocracyDecoded]

Tesla got more brains i feel, however they lack boston dynamic's robotic tech.

The arm movment for Boston dynamics robots are among the fastest and the most fluid I've seen.

[saiine]

This is like comparing a smart phone to a landline. Most folks missed the implications of putting FSD and training a neuronet for a bot, and I guess I understand that if you're not an engineer or paying close attention to what's happening.

[Jkim]

AI development is exceedingly hard. Making the chassis and shell is pitifully easy compared to programming the movement. Every complication, every additional task and development exponentially compounds the difficulty of development, and bipedal walking is probably one of the most complicated forms of locomotion one can ever create. Far more complex than wheels or having more than two legs.
Optimus was clearly designed for two legged motion in mind. It's supposed purpose is to create an affordable and widely accessible AI robot for completing...something? "General purpose tasks?" "Using tools?" Something that needs to have stable locomotion in most conditions and environments with a reasonable degree of strength and dexterity.
Every single complication creates an exponentially compounding affect on the difficulty of overcoming the next step. Just an empirical example here, if programming a robot to pick up a box, lift it, and set it on a shelf in front of it takes 1000 man hours to program, I wouldn't be surprised that getting it to do that onto a shelf off to its right without falling down would take 10x longer. Moving without falling down, doing that with two legs, in varying terrains, avoiding moving obstacles like people, understanding of speech, voice and facial recognition, doing physical tasks. WITH. HANDS.
My point is, stuff like this takes massive amounts of money and lots of time involving countless baby steps. That's why we've been at this for decades, culminating in Boston Dynamics' astonishing and praiseworthy accomplishments. They have for years now created capable AI that can navigate and do simple manual tasks with their own sensors and brain independent of any scripts. Any other businessman who wants to get into the market and do all this while creating something cheap and affordable has some avenues to take. Collaborating and mutually benefitting from the ideas and work of those who have already done it before. Will Elon Musk do that? Nope, start from scratch for...reasons. Okay, take a long time to try and get it done right because you want it built exactly the way you want.
"3-5 years."
Perhaps start with a much more expensive model that allows for it to perform within expected metric, and slowly decrease its price over time while your company works out various performance and manufacturing issues until you finally get something more accessible to the average consumer. Kind of like what Tesla already did with their electric vehicles.
"It will start at $20, 000."
Okay. Perhaps he's super talents on a miracle level and can really fulfill a major percentage of his projects within the stated performance metrics, budget, and prices of sale. How is his track record on fulfilling all of that with his proposed ideas?
*looks at SpaceX (just SO much of SpaceX), Starlink, Tesla 3, Tesla in India, Neuralink, Hyperloop(s), OPENAI, SELF-DRIVING AI, Tesla Semi-truck, Robotaxis, Cybertruck, Solar Roof, Boring Bricks, etc etc etc...*
Perhaps the fault isn't in the technology. Perhaps it's the man behind this Optimus project who raves about returning to the Moon and going to Mars. I bet that if this billionaire charlatan hadn't shown up, it would be a lot more likely we'd be much closer on our way to the Moon and the Mars.

[martinlutherbling]

Boston Dynamics design is LIGHT YEARS ahead of Tesla.

[harlowblackadder]

How does Elon get away with just straight up lying year after year after year?