this post was submitted on 11 Sep 2025
195 points (96.2% liked)
Technology
75238 readers
5142 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Over time, maintenance costs on machines tend to increase. They all have a practical limit on profitability, before that cost exceeds their contributive value. Then they need to be replaced.
And if the machines are the ones building new parts, that, like many other things, goes out the window. They can even recycle and refurbish parts
That's pure science fiction. It will never happen. Training people to do various manual tasks is always cheaper than using robots. Automation involves dedicated, task-specific machinery that improves on existing (manual) methods. People are always there to fill in the gaps in what those machines are capable of. We provide that required versatility.
Replacing people with people-shaped robots to do the exact same job that people do, is the opposite of efficiency. There is no improvement involved. It's literally a lateral shift, with an enormous price tag attached to it.
I don't know what to tell you, other than it's already happening. Once the first robot builds a second, it's over. You can buy one that can physically do light tasks for $8k, this summer Amazon started using robots for deliveries and has been using them for packaging for longer
It's not science fiction, it's now an engineering problem, one that is progressing quickly
Lol! Dude. It isn't "already happening". Where are you hearing that?
And are we still talking about humanoid robots, or are you talking about drones and automated roller carts? Because they do have those, but there's no way they are able to repair each other or build more of themselves. What they do have, is as I said, very task-specific and non-intuitive. If even one variable is out of place, the whole system goes off the rails, and an actual human being is required to put things right again.
No, I'm talking about automonous humanoid robots specifically. The rollers and shelf bots have been around for years
NVidia also just released a big suite of tools to train AI for robotics, it's basically a huge physics sandbox where you can train and test models at scale before real world testing
Boston dynamics and others are currently writing/lobbying regulations for bipedal robots so that they can meet safety requirements - current safety standards require an emergency shutoff switch, but bipedal robots fall over if they don't balance, which isn't particularly safe
This is happening, and quickly. None of them have the dexterity to machine parts, but the range of tasks they can do is rapidly expanding
Lol! This isn't "happening, and quickly". Boston Dynamics has been working on their humanoid robots for decades, and they're basically at the same stage they were at the beginning.
It's just a gimmick, my friend. Not a viable alternative to human labor. They don't perform tasks "better" or "more efficiently" than people. It isn't even a matter of them improving over time. You simply don't invest in new technologies that promise to do the exact same thing as the old ones.
Lmao... That's a wild take. Boston dynamics has been steadily improving this whole time, they were the first to really crack bipedal locamotion. Not just walking, running and flipping. Carrying loads. Kipping back up to their feet
You can, right now, for $8k buy a humanoid robot that can run, and be controlled to do whatever else. That's insane
And you can get shelf stacking humanoid robots that work commercially. These exist and are for sale
Amazon is currently field testing humanoid robots that deliver packages from the truck to the door.
Your knowledge is very dated, friend.
Lol! Are you talking about this? Dude, this is what I meant when I called them a gimmick. And if I recall correctly, the "shelf stacking humanoid robots that work commercially", are not actually"working commercially". In fact, they didn't work at all when given actual things to lift and stack. They could only carry empty boxes, and dropped them more often than not, and tended to fall over all the time.
Like I said, even if they improve to the point where they don't fuck everything up...all they will be able to do, is the same thing people already do. Except people can also do all kinds of different things, without requiring an engineer to be onsite to set them up for the new task.
No... I've never seen those before. That's not what I'm talking about at all, I also think the Tesla robots piloted by humans are probably going nowhere, for the record
Want to run failed startups past me some more? I gave you examples of humanoid robots being tested in real world conditions
Ummm, except you didn't give me examples. Can you post a link to these robots that are actually being tested in real world conditions?
Or even a link to the $8k robots that are not the same as the one I found for $6k? I have a hard time believing that another $2k is going to somehow provide the difference between that thing, and something functional.
Yeah ok
Amazon delivery robots
$8k robot
Lol! The Agility Robotics model used by Amazon costs $250k per unit. And is currently only capable of moving empty totes around in a controlled environment.
And the Pudu model you linked, is listed on their website for $214k.
I know I half assed those links, but your numbers are an order of magnitude off lmao
Were you not able to open the link to their website? You can buy one today for $214k. It's the D9 model from your article.
Did you not convert from yuan or something lol?
Buddy, it's right there on the page I linked. Even your article says they're hoping to get the price down to under $30k, at some point. That's why I looked up their website directly, to see how much they're currently selling them for. It's $214k. I don't know what else to tell you.
I see one for $43k on there, I don't see any for $200k on that site
But that's kinda irrelevant, as is how much Amazon's robots cost. The point is how quickly robotics is advancing
There's so many companies making these that I can't find the one I saw weeks ago. They're being tested in real world conditions. There's a million groups playing with these things, trying to get them to do more and more tasks
And think of what full automation would look like... It would be a bunch of factory equipment like it is now. The automatons just need to carry and place stuff between stations, maybe slot things together and screw them in. That's not that high a bar
I just don't see the line where any of that is impossible. It seems inevitable if society doesn't collapse this decade
The disconnect is coming here:
I agree with you that robots building robots is not impossible. I disagree with you that it's therefore inevitable. I strongly disagree with you that it's therefore inevitable in the immediate future.
I think it's inevitable because capitalism, capitalism loves to treat people like replaceable cogs.
These would be literal replaceable cogs that never rest, don't have rights, don't need to account for safety concerns, and don't have any of the other messy human bits
And even if we managed to suddenly outgrow capitalism, being able to have robots making robots in space would be a huge game changer
I don't think you're arguing in good faith,at this point. You send me an article about a specific humanoid robot, claiming it costs $8k. That article clearly states the make and model, as well as provides pictures of the unit in question. It is the Pudu D9 humanoid robot.
So, I sent you the website where it is being sold, which clearly states a price, that doesn't match your claim...and now all of a sudden, you don't even know what robot you were talking about before?
Give me a break, man.
Yeah, well, searching things is impossible these days, and after clicking around for a while I found one that looked similar.
I did read an article about one, went to their page and it was listed as $8k, I looked up the specs and it looked like it could pick up about 20lbs and has a weak but reasonable grip, and it made an impact. Then I saw video of someone recording one running down a sidewalk - it was a cell phone recording, but the robot was controlled by someone filming a demo in public
I don't really care if you take me at my word or not, the price factor to me means more because it means every robotics program in the world will be able to play with them.
But again, this isn't the core of my argument at all, but by the time I was looking up links I was kinda getting bored with this. I like to argue over ideas, it's a field I'm following closely with so much happening and we're just too far apart to make this constructive.