The Spot dog (which inspired the Black Mirror "Metalhead" episode) in 2016 https://www.youtube.com/watch?v=tf7IEVTDjng
Atlas doing backflips in 2021 https://www.youtube.com/watch?v=FByY3tSx2Ak
So 5 years of progress within a year.
Boston Dynamics is the follower here.
I know very little about robotics, but given these appear totally free-standing, if that was the case (I personally don't think it is), wouldn't that imply they have the same centre of gravity and weight of limbs as humans? Surely they'd have to be able to balance themselves, and copying a human's movements "exactly" wouldn't work for their own motion otherwise?
I think when watching I saw one or two of the robots "judder" their feet a bit out of sync with others - this seems to imply they are capable of balancing their own motion a bit individually.
These robots are certainly running through a scripted set of poses which has been extensively tested for the conditions (Humans would also be choreographed and have to hit certain marks at certain times). If you covered the stage in loose gravel or a thick carpet they'd all start falling over. The things the robots hold are almost certainly taped into their hands.
Despite that, this is a very impressive demo. Those robots are $40k+, they've got 20+ of them. And not a single one fell over. They're fast too - and there are a load of corners they could have cut, but they didn't.
The floor has two textures, it would have been easier without that. The humans right alongside them? Much less safety paperwork without them. The robot wearing trousers and a cape? Much easier without that. The fewer robots you have, the lower the chances on falls over landing their backflip. Lose the audience and record it in multiple takes. Hell, you could have human acrobats in robot costumes and it'd cost far less and be much easier.
So this demo is very much a costly signal of confidence.
You can clearly see that the robots change their grip of their sword, so it cannot be taped to their hands.
The motion is certainly scripted, but the exact mechanics in play there almost certainly vary radically from take to take. Imagine something simple like a pool/billiards break. Even if you set up a machine to rack the balls and break them in as close to identical as possible, you'd get wildly different results each time. And the dynamics in this motion is going to dwarf that.
That was of of the two things that impressed me most, along with the choreography involving close and direct contact
Not a fan of bipedal platforms or 50kg of servos for a number of reasons.
Best regards. =3
Does anyone remember when Honda's Asimo robot clumsily fell down the stairs during a demonstration[1] and we thought we were safe from a robot invasion by just moving to the upper floor? That was about 20 years ago.
The hard part with “autonomy” is interpretation of the environment and feeding that back into some control loop to accomplish a goal in real time. That is why most of these demos are basically recordings of movements, like choreography.
check this 4 months old video below
https://www.youtube.com/watch?v=bPSLMX_V38E
I'd willing to bet that it is already close to impossible to get the robot lose its balance without some significant external forces.
https://www.youtube.com/watch?v=jXTibM33SDg
However, then another short video bit alike popped up and is puzzling too.
Apparently Unitree robot is playing pingpong match like a pro. Sorry about german announcer, I couldn't find with english.
https://www.youtube.com/shorts/BgD1ukTyNnw
There is another match viewable by pressing that "Robot plays ping ppng #robot" arrow.
How about that robot? Is it human assisted or not? Our opinions diverted, I'm quite sure it is assisted but my former colleague thinks it's got to be autonomous as it would be too difficult and slow to do that fast movements with remote control assisted robot.
It would be nice to hear opinions about that playing robot too if anyone could provide some insight in that.
edit: I think the serve waiting robot hand movement and after losing wiping left eye gesture as a disappointing a bit in my opinion gives up it's human. Or if not, why would a robot do such a human like gestures.
edit2: OK, good points, I see now. It's definitely a fake. Thanks to all who replied :)
The robot is floating above the ground.
The paddle is phasing in and out of existence.
The robot has a realistic human hand and uses it to hit the ball.
The robot randomly turns around mid-air near the end of the video.
The robot looks nothing like a Unitree robot.
Oh, how could I forget, the entire robot looks so obviously fake even when disregarding all of the above that I can't believe you're even trying to analyze anything in that video.
2) Any semblance of American technological superiority is pure fantasy at this point. The only area where Americans are truly "advanced" is in selling overpriced SaaS products. There are dozens of Chinese startups with robots just like this—as seen at CES—yet Boston Dynamics is still treated like it’s some untouchable, DARPA-level tech.
3) A lot of this comes down to cost: you can either hire one American fresh grad or a Chinese PhD for the same price.
3) The second reason is cultural: Americans tend to buy solutions, while the Chinese prefer to build them. Even SMEs in China maintain internal dev teams to build custom software for the business, as opposed to paying Salesforce for what is essentially a glorified Excel sheet with sprinkles of automation.
4) America is facing its own innovator's dilemma. The country is currently being run by MBAs and salespeople focused on extracting every last dollar from the consumer instead of providing real value or innovating. Perhaps we're one step beyond the innovators dilemma. The innovators are dead and we are in the corporate greed stage.
5) Americans are completely oblivious to how advanced China has become because of the propaganda they're fed. My personal "aha" moment was when Chinese EVs hit my local market and completely obliterated legacy automakers on both features and price. The American "free" (lol) market is being guarded by politicians but that won't work for long.
Money should be a means to achieve a thing, not the goal in and of itself. I think the most visible decline came with the increasingly overt goal to charge rent on friggin everything. That's simply not a sustainable or realistic economic model for society and consequently even if it might maximize corporate income in the short to mid-term, in the longterm it's equally catastrophic for them as well.
Every Jane Street hire could be building robots, but instead, they’re trading options and crypto and heck, even market making for prediction markets now
there was this Chinese company named Baofeng that built a stupid media player by "re-using" open source FFmpeg code, it managed to get itself publicly listed, then had its valuation went up like 50x for doing nothing other being accused for stealing FFmpeg code.
there were lots of discussions at that time how that happened and why the same level of speculation didn't happen on other public tech companies listed on Chinese market, the consensus was pretty sad - tech companies suitable for speculation are listed in the US by default, those listed on Chinese markets are 2nd tier or 3rd tier to start with, they don't offer any meaningful room for speculation.
Crypto is a good example of how the equilibria is hard to maintain, and if the last cycle saw many interesting new products come to life, they all got crushed by ruthless profit-taking from early investors and team members.
On one hand Boston Dynamics showed similar skill robot well before this demo, only without coreography, which is were most of the wow effect comes from here.
https://m.youtube.com/watch?v=UNorxwlZlFk
Heck check were they were 5 years ago
https://m.youtube.com/watch?v=fn3KWM1kuAw&pp=0gcJCUABo7VqN5t...
Things is american research is financed by outcome potential not for grandstanding, and free standing robot that can only do recorded coreography aren't that useful outside factory floors, and factory floors can use ceiling rails or wheels to better effect.
So yeah video is suler cool, but there isn't much to it beyond that to read in terms of capabilities. You seem just to be projecting what the truth you want to be on top of a funny dance.
Meanwhile, the US installed 34,200, a decrease on the previous year, and virtually all of those were imported.
https://ifr.org/ifr-press-releases/news/global-robot-demand-...
Ah yes that explains Tesla...
What impressed me the most is the amount of EVs on the streets.
Imagine you need weeks to start a new software module development and to procure cloud instances.
After a while, naturally the locals would buy the white label products that are anyway the same as the branded ones, many times produced on the same factory lines.
My father used to say, every company goes downhill when management takes over, meaning those straight out management schools without any actual business experience on what the company does, and he was kind of right, that is how we hand landed in late stage capitalism and entshitification, in the middle of geopolitics turn over.
These robots might not drive the car for us, but certainly will become part of some police containment unit, regardless if they are remote controlled or AI driven.
Once China built an industrial base, and has cheap energy sources, you cannot directly out compete it. You can only try to maintain your own industrial base by locking competitors out of your market. There is no other way. In the end that's the result of globalization - US&EU companies thought they can produce cheap, sell expensive. Instead they trained their own competition, and due to weak IP laws enabled this exodus of industry know-how to China.
Now, if China was a smaller country - let's say Japan or Turkey - this wouldn't be such a huge problem. But for the global economy, having a single country that produces 80% of all consumer goods is also a huge problem. That was never the case before with the US (except maybe directly after WW2). US+Europe, Japan+Korea, Canada, Australia supply chain was much more diverse and distributed.
What happens now is dangerous because in the end the profits are not spread across the world, and economies of scale cause this monopoly to appear, which will be hard to mitigate.
Can countries "slow" down China and move production to diversified locations? For that to work, coordinated tariffs for advanced goods from China would have to be introduced, and production reallocated to multiple other countries - very difficult to execute..
They'd only have been 10 in 1999 when cell phones were pretty ubiquitous among adults. I'd say they were basically past it already.
I was born ten years earlier and I'd agree that the internet is likely the biggest change in my lifespan, but I'm not sure I'd say it was as sudden as the past 2 years of AI.
If I look at the most useful technologies around me, then Google Maps ranks high. But it wasn't sudden. It was on the desktop first. And then slowly crept through mobile.
LLMs on the other hand, suddenly just kind of appeared.
They are very different robots with very different goals, so it should be no surprise that the G1 appears much more agile.
some specs here: https://www.unitree.com/H2
claimed 3h battery life, can hold about 10% of its weight (7kg, with arms)
I personally agree with the rest, recommend you remove that last bit about downvoting and sneering if you can still edit
You can see on the backflips that all robots landed quite differently, some with both knees on the ground, some with one, some with none. Yet all recovered gracefully and moved on to the next step of the choreography.
It is genuinely impressive, and scary.
Meanwhile in the west we are bickering like 10year olds.
Unitree's army of robots
https://www.youtube.com/watch?v=w4IOJH9Akhg
Robotera sword dancing robots
https://www.youtube.com/watch?v=9Ti9Mi8rbIQ
AGIbot's flying kicks
https://www.youtube.com/watch?v=QXnXdh6IEkA
LimX's Tron2 robot
https://www.youtube.com/watch?v=Ut3QFPr7hyo
one interesting observation is that none of those companies are located in Shenzhen, which arguably has the best supply chain for all electronics stuff. I guess those trillions $ spent on infrastructure paid off - Shenzhen didn't suck all talents into its proximity, it becomes an enabler for industries across the country.
AGIBot is the poster boy of the municipal government of Shanghai.
> major cities/provinces like Hangzhou, Suzhou, Shanghai or Hefei started bankrolling talents and enterprises so they too can have major high tech enterprises in their town
State Capitalism at its best.
We are definitely on an exponential in term of capabilities of humanoid robots. We are probably only years away from having a robot in the house, in construction of robots. Automating anything that a human can do is best done in a human sized robot.
But.
None of these are actually useful right now. I don't want something with the arm strength of a forklift taking care of my parents or kids. The demand for humanoid robots right now is like lift a fridge from a delivery truck to a house (aka more mobile forklift) or walk through toxic sewage to pull crates out. Super useful but basically just mobile cranes, which is a small market. China seems to be making the mistake of pushing a tech demo as a consumer product (we've all been on those projects...) which can make people hate the tech.
Build something people want, don't mandate what they want. We're like 3-4 generations from amazing, useful robots. I'll be scared when these things are minding a bunch of dogs on stage.
They could be really useful: without hesitation such humanoid could bring pack of explosives to the opposed treeline.
Also, current tech could be useful as a shopping assistant, to carry the groceries for people who can't, for one reason or another. Though the other post about tipping safety does have a point.
A risk I never hear discussed is falling over and injuring children. Even the petite Unitree models are like a 70kg piece of furniture. Each year thousands get injured because of furniture falling over. I'd buy one immediately, but if I had kids or pets, I would wait for safety data on falls.
Robots in such an environment are designed with the appropriate affordances so that they cannot use too much force... but the concern about weight I suppose is quite salient.
You had to use a few trick for larger programs, but basically it managed to type any real programs and I never encountered an ambihuous case that caused a problem in practice. In case of failure, the small set of unsatisfiable constraints was easy to translate into nice error messages. This also allows for typing rules that are easy to state and can accomodate operators that adapts nicely to their environment.
I would understand if this approach would be frowned upon, but I still wonder if any serious language ever used this approach?
Was intended for https://news.ycombinator.com/item?id=47025885
And z4 is, of course z3 (no I don't have special access to next-gen version of z3)
there is reason why most of the shots are not wide angle showing whole scene, seems they learned their lesson from last year where you could easily see on the edges all the failures
this was heavily edited and repeated, I mean is it really surprising considering all CGI you see during whole gala? I watched whole 5 hours (though skimmed through a lot), they just can't make show same as seen by real people on the site, what you see in TV is very different from what audience has seen
edit: the whole gala show is recording, it would be impossible to organize such event across many cities with so many performances live, olympics opening ceremony is walk in the park compared to this
Reason: otherwise, you will be asking what happens when the shot is switched between the stage and the audience.
While I don't know whether this was indeed broadcast live, at least this recording is missing a section since as the YT comments point out at 1:25 the staffs appear out of nowhere.
I was initially somewhat skeptical as well because this looks like a surprisingly massive leap in robotics capability, but haven't been able to find anything particularly sus.
Now the nice thing about demographics is that it's eminently predictable. IMO this is why China is going all in on robotics.
No less impressive, but is it likely each robot autonomously learned a routine? Or just got programmed for a very exact act?
