But the prototypes used "artificial intelligence" to control their walking.
So the thing that Asimo and Boston Dynamics have been doing better and more fluidly for years if not decades, that's what was being driven by the thing Tesla has staked its financial future on and scheduled this entire event to impress people with. Everything else had all the sophistication of the Pirates of the Caribbean ride at Disneyland, or was being controlled remotely.That lines up with reporting from tech blogger Robert Scoble, who posted on social media that he had "talked with an engineer" who confirmed that "when it walked, that is AI running Optimus."
Funny enough, Marques Brownlee made the same comment (though he was commenting about the taxi).Sounds like these were more "concepts of a plan" for robots than actual robots.
Stomping along on perfectly flat ground, with no variation in gait or cadence, with all observers kept far enough away that nobody could fuck around and push one.For how long?
Also, what is their definition of "walking"? Experience with some IT types has taught me some people use such insane, narrow, non-standard definitions that they tend to be close to barebones lies.
Funny enough, Marques Brownlee made the same comment (though he was commenting about the taxi).
View: https://youtu.be/fgm5uZaS3-E?si=sCoyJs62p2u7F-76
It was so abysmally obvious just from watching the video that these things were being remotely operated that I genuinely question whether this Munster dipshit has enough on the ball to manage his own bank account, let alone Deepwater Asset Management. I wouldn't let this guy manage my tickets from a game of ski-ball at Chuck E Cheese, lest he invest them all in the Mega AI Use Case known as Munch's Make Believe Band."Fooled me," Deepwater Asset Management Managing Partner Gene Munster admitted on social media after hearing reports of Optimus' teleoperation. That admission came just hours after Munster posted about how the event was "just the start of mega AI use cases."
Not very except they are very litigious.How different is this from the whole Nikola rolling their Semi downhill debacle?
Are we now at the point where Proportional-Integral-Derivative control systems are called "AI," even though they've been around for over a century?The report doesn't specify which demonstrated capabilities needed that human assistance, but it points out that the robots "were able to walk without external control using artificial intelligence"
He's just "crowdsourcing" using already existing infrastructureSo what you're saying is that Elon Musk trained a neural network* to operate Optimus? That's amazing, time to hodl TSLA!
* using off-the-shelf hardware not manufactured by NVIDIA or TSMC!
Just the people paid hundreds of thousands to manage other people's hundreds of millions, I guess. Because apparently Gene Munster and Cathie Wood bring that good galaxy-brain shit the rest of us don't, or something.Was any one really surprised by this?
And none of it puts them even close to the competition, let alone ahead of them like they desperately want rubes to believeSo, hardware is looking reasonably polished. Has low level autonomy (self balancing, walking, etc) but the software to "do stuff" isn't demo-ready, especially in a chaotic scene like a party (vs. factory floor ... behind a fence).
That's all fine.
Using teleoperation in the meantime makes for a cool demo, while the software gains features and matures. But JFC, be honest about what you're showing!
There should be no shame about a demo not being final features, about stuff not being ready. But you've got to be honest about what's "real", and what's a stand-in to help people imagine possibilities.
Musk very pointedly avoided discussing the autonomy of the current Optimus prototypes during his "We, Robot" remarks. Instead, he simply pointed out that "the Optimus robots will walk among you... I mean, it’s a wild experience just to have humanoid robots, and they’re there, just in front of you."
I suspect the presentation was tailored to imply as heavily as possible without outright lying. Thus the whole being coy thing about what part was AI. And the forward looking statements (ie, we will be able to do X at some point, which can't really be proven wrong, vs we can do X now - when they can't).So, is the FTC or the SEC the ones that have to look at the full presentation to see if there was any violation of federal rules?
FTC handles false advertising and the SEC misleading of investors, right?
No, they're not even close.They've basically reached feature parity with other robots like Asimo and offerings from Boston Dynamics.
I was fairly surprised how much he called them out... Just to end with "But if any company can do it, it's Tesla" ...He's still so fucking desperate to avoid saying anything genuinely critical of Tesla that it trashes the objectivity of anything else he might review.
Hardware is the easy part and even at that, this is decades behind the curve. And as with every complex project, the Pareto Principle reigns; they're not even testing the boundaries of the 80% that takes 20% of the effort yet. They might be breezing through the initial 30%.So, hardware is looking reasonably polished. Has low level autonomy (self balancing, walking, etc) but the software to "do stuff" isn't demo-ready, especially in a chaotic scene like a party (vs. factory floor ... behind a fence).
That's all fine.
A demo not reflecting final features or not being fully ready for release is one thing. This is quite another; it wasn't a demo, it was straight up science fiction. There was nothing being demonstrated. It was like taping cardboard to a Sopwith Camel to "demo" a god damn F-15 Strike Eagle. It's an outright fraud to represent Optimus as even notionally or theoretically capable of what Tesla portrayed, especially without being crystal clear that they were all teleoperated.Using teleoperation in the meantime makes for a cool demo, while the software gains features and matures. But JFC, be honest about what you're showing!
There should be no shame about a demo not being final features, about stuff not being ready. But you've got to be honest about what's "real", and what's a stand-in to help people imagine possibilities.
They're so far from feature parity with Boston Dynamics they're not even playing the same game, homie. Boston Dynamics was at this point 15 years ago, if not 20; they're doing parkour and shit now. Best believe that we got as much as Optimus can actually manage; if a dozen Optimi could sprint out and put on a flashmob Bollywood dance number while Elon stepped out of a Cybercab that'd driven itself through city streets to deliver him, that's what we'd have seen. And we sure as fuck did not.They've basically reached feature parity with other robots like Asimo and offerings from Boston Dynamics. The big problem they have is the next step is basically "develop AGI", which Elon has been absurdly bullish on solving with machine learning despite the seemingly stalled progress on that front. It's also why he seems to think all he needs is a bigger datacenter to solve driverless cars. Maybe it will work out, but right now the signs aren't promising. There is a very noticeable degree of diminishing returns as the models get bigger and bigger and the last 10% of the problem is taking more than 90% of the effort.
You forgot your '/s'Maybe it would be more accurate to say that Tesla is the only company really trying to solve this problem. If they don't do it there is nobody else in the field of general purpose autonomous humanoid robots. Most people don't think it will even be practical without some major breakthrough in the field of AGI.
Maybe it would be more accurate to say that Tesla is the only company really trying to solve this problem. If they don't do it there is nobody else in the field of general purpose autonomous humanoid robots. Most people don't think it will even be practical without some major breakthrough in the field of AGI.