arthwollipot
Observer of Phenomena, Pronouns: he/him
We have humanoid robots because we want humanoid robots, not because a humanoid is the best platform for most applications.
Cont: Musk, SpaceX and future of Tesla II
- Thread starter a_unique_person
- Start date
jeremyp
Philosopher
Yes I was thinking about Boston Dynamics when I made my post. I watched a YouTube video about how they developed robots for loading and unloading delivery trucks. They started with something that looked vaguely like a human but then realised it was a poor shape for the job and iterated the design several times, ending up with something that looked very different.
Can't find the original video, but this one shows the end point.
jeremyp
Philosopher
Do we really? Or is it just sci-fi enthusiasts and tech-bros and ketamine loaded billionaires seeking to shore up the share price of their company?
theprestige
Penultimate Amazing
Cherry picking. Boston Dynamics is also working on the Atlas, a humanoid model.
theprestige
Penultimate Amazing
Boston Dynamics originally came to fame with its prototype quadrupedal robots. The tech they developed there went into things like the quadrupedal Spot, and the humanoid Atlas. I'm sure some of it has also made it into their waldo-arm design.
Look, Elon Musk is dumb about a lot of things. But just because he's dumb about pickup trucks, that doesn't mean pickup trucks are a bad idea. Just because he's enthusiastic about his Temu Atlas doesn't mean more capable robotics companies like Boston Dynamics are on the wrong track.
JayUtah
Penultimate Amazing
I can think of some applications for anthropoid robots. In the practical sense, if you have environments originally designed for humans but in which it is no longer desirable or safe for humans to work, you might want an android to navigate and interact with the environment. You can have androids augment human labor, where the higher-function tasks are done by humans and the dangerous, repetitive, or heavy work is done by androids. I can think of some entertainment concepts too. One of my former customers, Sarcos (now Palladyne), built a lot of Disney's more realistic animatronics such as Tom Morrow at Innoventions and Capt. Jack Sparrow. I think they sold off their hardware division to Raytheon to develop combat exoskeletons, so keep that in mind next time you ride Pirates of the Caribbean.
Boston Dynamics' Atlas isn't being developed for any given market application, as far as I know. It's a research and development platform. This implies that technologies they're developing there are useful in more market-driven applications such as the quadrupedal platforms. Without thinking about this very hard, what immediately leaps to mind—pun intended—is advanced locomotion. Being able to run over uneven terrain would be useful in a search-and-rescue scenario, or the security scenarios they already contemplate.
All the robots I use are waldo-arm types, most of them either bolted to the floor or desktop models bolted to breadboards. And this is because I need them for precision movement and placement. If you see, for example, circuit-board population, that's all just a specialized robot arm. There's no reason why "robot" has to be exactly a waldo arm, or exacly and android, or exactly a terror dog. Consequently there's nothing that says a surgical robot has to follow any of those forms and can't be designed to solve the special problems that surgery presents.
Now as far as replacing all human surgeons with robot surgeous in five years goes, you know what they say: prediction is hard, especially when it's about the future.
Boston Dynamics' Atlas isn't being developed for any given market application, as far as I know. It's a research and development platform. This implies that technologies they're developing there are useful in more market-driven applications such as the quadrupedal platforms. Without thinking about this very hard, what immediately leaps to mind—pun intended—is advanced locomotion. Being able to run over uneven terrain would be useful in a search-and-rescue scenario, or the security scenarios they already contemplate.
All the robots I use are waldo-arm types, most of them either bolted to the floor or desktop models bolted to breadboards. And this is because I need them for precision movement and placement. If you see, for example, circuit-board population, that's all just a specialized robot arm. There's no reason why "robot" has to be exactly a waldo arm, or exacly and android, or exactly a terror dog. Consequently there's nothing that says a surgical robot has to follow any of those forms and can't be designed to solve the special problems that surgery presents.
Now as far as replacing all human surgeons with robot surgeous in five years goes, you know what they say: prediction is hard, especially when it's about the future.
jeremyp
Philosopher
Yes, but Atlas isn't for anything except research and making exciting videos. Look at the stuff they build for real applications that people will pay for. They don't look like humans and humanoid robots will not be able to compete with them.
SteveAitch
Master Poster
Thinking about, and I am not an expert, I would have thought a four-limbed, sort-of humanoid robot would have some advantages; in exploration. for example. Fine upright on flat or mildly sloping surfaces, able scramble up steeper surfaces and to climb (like a human?) much steeper inclines that a tracked/wheeled vehicle would have problems with. Being able to move its centre of gravity would be an advantage. I think.
Mind you, I am the person who thought up the egg-shaped unmanned lunar-lander based on Weeble technology...
Mind you, I am the person who thought up the egg-shaped unmanned lunar-lander based on Weeble technology...
Last edited:
theprestige
Penultimate Amazing
I disagree. A lot of the built environment is scaled to human forms and optimized for human ergonomics. A humanoid robot that can move and work in human-centric environments will be valuable.
dirtywick
Penultimate Amazing
- Joined
- Sep 12, 2006
- Messages
- 10,063
as it applies to manufacturing, most people think of robots as arms that articulate on a few joints. and there’s applications for that to be sure. but most automation is done by a machine built for that particular process. a robot that looks like a guy might be useful in multiple applications, but usually you buy a machine that does one thing in particular really well, and those usually don’t look like a guy.
Andy_Ross
Penultimate Amazing
- Joined
- Jun 2, 2010
- Messages
- 66,707
jeremyp
Philosopher
Would you agree that delivery trucks and warehouses are scaled to human forms. Now watch the video I posted.
jeremyp
Philosopher
Yes, because it is a lot easier and cheaper to do that. We don't yet have useful humanoid robots because they are very expensive and hard to make do things well. But, even in my brother's house, there is a robot for doing the dishes, another for doing the laundry, one for vacuuming the carpets. They just don't look like people.
Think about Waymo taxis: are they driven by robots sitting in the driver's seat? No, because it's much easier to have a motor connected directly to the steering wheel and another to the gearbox and fuel injection than to teach a robot to hold the wheel and use the pedals. I don't think any humanoid robot is yet capable of even getting into the car - that's quite a complex operation.
Here's my prediction: for any given task that can be automated, a specialised machine will always be better and cheaper than a humanoid robot.
JayUtah
Penultimate Amazing
Emily Glazer is reporting for the Wall Street Journal that the Tesla board has officially opened the search to replace Elon Musk as CEO.
angrysoba
Philosophile
Cherry picking. Boston Dynamics is also working on non-humanoid models.
arthwollipot
Observer of Phenomena, Pronouns: he/him
It's paywalled, but we can get the gist from just the title. Interesting word choice - "succeed" rather than "replace".
jeremyp
Philosopher
One report I watched on YouTube suggested that this was potentially looking for somebody for when Musk eventually steps down, hence "successors".
The Guardian, on the other hand, is a running a story that the whole thing is a lie and the board is denying all knowledge.
Tesla denies report claiming board looked to replace Elon Musk
Wall Street Journal article saying headhunters were contacted is ‘absolutely false’, says company chair
www.theguardian.com
Robyn Denholm says:
It's delightful how she says "continue executing on the exciting growth plan" as if she thinks Musk is executing on it now. It's a fact that Tesla has not grown in 2024 so how can he be executing a growth plan?
SteveAitch
Master Poster
Depends on what you mean by 'execute'...
Growth can be negative as well as positive...One report I watched on YouTube suggested that this was potentially looking for somebody for when Musk eventually steps down, hence "successors".
The Guardian, on the other hand, is a running a story that the whole thing is a lie and the board is denying all knowledge.
Tesla denies report claiming board looked to replace Elon Musk
Wall Street Journal article saying headhunters were contacted is ‘absolutely false’, says company chair
Robyn Denholm says:
It's delightful how she says "continue executing on the exciting growth plan" as if she thinks Musk is executing on it now. It's a fact that Tesla has not grown in 2024 so how can he be executing a growth plan?
