They are making good progress, which is nice to see. As someone who worked with the early version of the Boston Dynamics humanoid robot, in the 10 years since this looks like a nice upgrade.
However, I am still not sure what Tesla wants the robot to do. The reality that walking robots are extremely inefficient and use a lot of power, since they don't walk like humans do. Especially on perfectly flat floors like a factory where wheels would be infinitely more efficient.
The humanoid arms and hands are very useful. Many tasks still require dexterity of a human.
However, at the end of the day, it is the human cognition and hand-eye coordination that is key. Without autonomy software to do that, these kinds of humanoid robots are just a waste of money for most things.
Even if Tesla could make them for $20k. What would you even do with them? They are just expensive toys at the moment to make Tesla look innovative. Musk is a hype machine, and this just feels like a gimmick to me. Maybe there is a business plan here, I just wish we knew what it was.
Legs are not the reason humans cant be replaced… most factories are single story due to the massive imposed loads from machinery, and even those that aren’t you could simply have robots assigned by section as tasks would tend to be dissimilar between sections anyhow. All of that could run on wheels no problem. So designing a humanoid robot serves no purpose except an overcomplicated design (there’s a reason humans spend a year learning how to walk).
The only case for bi or quadrupedal robots I’ve heard of is for traversal of natural terrain, which is really only useful in a military or SAR application.
So yea, call me a skeptic, but imo any robot on legs where the legs serve no definite purpose explained by the manufacturer is a marketing gimmick, not a real product with an actual use case.
They are making good progress, which is nice to see. As someone who worked with the early version of the Boston Dynamics humanoid robot, in the 10 years since this looks like a nice upgrade.
However, I am still not sure what Tesla wants the robot to do. The reality that walking robots are extremely inefficient and use a lot of power, since they don't walk like humans do. Especially on perfectly flat floors like a factory where wheels would be infinitely more efficient.
The humanoid arms and hands are very useful. Many tasks still require dexterity of a human.
However, at the end of the day, it is the human cognition and hand-eye coordination that is key. Without autonomy software to do that, these kinds of humanoid robots are just a waste of money for most things.
Even if Tesla could make them for $20k. What would you even do with them? They are just expensive toys at the moment to make Tesla look innovative. Musk is a hype machine, and this just feels like a gimmick to me. Maybe there is a business plan here, I just wish we knew what it was.
I think that's funny. You really can't imagine what a human might do, but prefer not to? How about sorting, washing, and drying laundry, ironing it, folding it, and then taking it up stairs to put it away in the kids' rooms? How about making dinner? How about running a CNC mill, which requires someone stand there by the door, and when the machine is done, then the operator (normally a human) opens the door, blows off the parts, loosens the vice(s), takes out the finished parts and blows them off, then replaces those finished parts with new blank pieces of metal, plastic, or whatever the parts are made of, and then tightens the vice(s), closes the door, and pushes the green button? Does that sound like a boring job to you? With a few Optimus robots to help me I could run five or six CNC mills, rather than the two I normally run at work. That would make me much more efficient, and the Optimus robots would pay for themselves in a couple of years . . . and then make the shop a lot of extra profit.
The thing is, Optimus will have a thousand uses. Tesla will make money with them not only by selling the robots (and servicing them in years to come), but they will sell upgrade packs (software/firmware), which will allow the robots to do stuff that it can't do as it comes. It will be like plug-ins for Photoshop or applications for a computer. You pay extra, and you get an upgrade pack that you can load into your Optimus, and all of a sudden it will know a thousand languages, or it will be able to fix your car, or it will be able to cook you a gourmet meal. The idea of the upgrade packs will be that you don't have to teach the robot how to do it yourself. (You might not even know how to do it in the first place, so you can get your robot to do it just by upgrading it for a fee.) What do you want it to do? There's an upgrade/plugin for that. (Imagine if you will . . . that Optimus is Neo, and you're in the Matrix - "I need to know Kung Fu.")
I can imagine Tesla making more money on the upgrade software/firmware than on the robots themselves. There will be thousands of upgrades available some day. A language pack, a chef pack, a mechanic pack, a maid pack, a helicopter pilot pack! Need a personal coach for golf? Download the golf pro pack, and upgrade Optimus to have a player to play with at the golf course tomorrow . . . but it will be one of the best players ever, because it was trained by Tiger Woods . . . and a dozen other famous golfers . . . and it has been trained to be a coach too, by three of the best golf coaches in the World. Sure, the golf pro pack costs $2,995. It's worth it though, because it will go golfing with you whenever you want, and it will have more patience than any golf pro or coach ever had. It has a library of 842 golf jokes too . . . just in case you're in the mood for some humor.
I believe that eventually Tesla will have more software people working on the Optimus program than Apple has working on Siri (and all the other software Apple makes). In fact, I believe that Apple is probably already working on their own robot. Humanoid robots like Optimus will be all the rage in a few years, and like smartphones, almost everyone is going to want one.