Between the Boston Dynamics robots that continue to perform more advanced stunts each year, and Tesla’s push into robotics with Optimus, it seems like this is the next frontier for manufacturing and labor. I wouldn’t be surprised if these kinds of robots were walking around on the streets by 2030.
To me, the obvious applications for this type of humanoid robot is something like an assembly line, where it is responsible for doing a finite number of well defined tasks with high levels of repetition. I’d imagine AI developments will make it so that these robots can perform more complex tasks like cooking, cleaning, and package delivery too.
A few questions for stackers…
  • What other tasks might these kinds of robots be better than humans at? (bonus sats for creative/interesting answers)
  • What will humans always be better at?
  • If you had an army of these robots at your disposal, how would your life change?
298 sats \ 3 replies \ @kr OP 26 Jan
i’m going off on a tangent here, but i’m also curious why all these Boston Dynamics robots walk about like they just shit their pants…
Why can’t they stand up straight like humans?
reply
It's called the center of mass :)
reply
42 sats \ 1 reply \ @kr OP 26 Jan
so is the issue that the hardware is too heavy/placed in weird locations on the body?
hopefully they can sort that one out before these robots are released to the public, they look very uncomfortable haha
reply
I'm pretty sure that's it, but I could be wrong.
reply
I think if we have to have robots, they should look like robots, not humans. Like TARS from Interstellar
I often imagine having one or two of these beside me at work, helping with heavy lifting, repetitive tasks, 3-D printing stuff for me, helping diagnose things. Basically the ultimate swiss-army-knife.
A tool should look like a tool imo.
reply
21 sats \ 1 reply \ @kr OP 26 Jan
interesting.
what kind of leverage do you think you could get on your time with 1-2 robots working alongside you?
would they be more or less useful than having 1-2 humans helping you?
reply
It's weird to think about. Like at what part of the curve of AI learning does it become more efficient to hire one Jack-of-all-trades human and 1-3 bots they they work with, program, debug, prompt etc. And then how much farther up the curve do I become obsolete? (like two months after??)
I work in research and to have 1-2 bots doing tasks 24/7, double-checking each-other, they don't get hungry, or sleep, or need time off, don't need benefits etc. would be pretty crazy.
I don't know, maybe AI will just blow past the whole robots in the workplace thing, because we won't need to work anymore, and we can jump straight to making our own open source R2D2's and TARS.
reply
Robotic warehouses.
Amazon is STILL one of the only companies to utilize fully-automated warehouses. And it has made them BILLIONS.
Expect this model to be replicated and exported to basically every warehouse in the world over the next 10-20 years.
Additionally, warehouses used to be where shelves held products until they were sold. But thanks to 3d printing, soon shelves will be making (growing) the products the instant they are ordered.
Humanoid robots are best for operation in human environments. Alongside humans, or in spaces where humans used to occupy. If humans need not apply, then the space can be adapted to be more efficient for a bespoke robot, at the cost of becoming a hostile environment for humans.
I doubt we'll see humanoids roaming the streets unattended anytime soon. Humans became the dominant species by murdering every other ape-like creature we encountered. Imagine drunk people throwing their electric scooter rental off a bridge but its a erie-uncanny-valley robot that you can bully around instead.
Robot designers will probably optimize for cute robots or boring R2-D2-looking trashcans rather than "human-like" robots as a defense mechanism for deployments that roam the streets.
reply
lots of good points here
Amazon is STILL one of the only companies to utilize fully-automated warehouses. And it has made them BILLIONS.
if Amazon is making so much money from their warehouses, why aren’t competitors building their own fully-automated warehouses?
thanks to 3d printing, soon shelves will be making (growing) the products the instant they are ordered.
which happens first? 3d printing at scale or warehouse robots at scale?
Humanoid robots are best for operation in human environments. Alongside humans, or in spaces where humans used to occupy. If humans need not apply, then the space can be adapted to be more efficient for a bespoke robot
good point, maybe the human form isn’t an optimal design here
reply
Amazon lost money for decades investing/researching/iterating on automation and it's not like they're publishing the blueprints for their competition to read. Walmart has only just started investing more in automation after it became clear that Amazon's bet was starting to pay off.
Many industries are looking at automation to supplement a part of their processes, but few are looking seriously at how to re-structure their process to be fully automated.
The buildout of an automated plant takes many years and requires many engineers. America is experiencing an engineering shortage. Additionally, most chips, electronics, and robot parts come from Asia which adds even more to the cost of building these kinds of deployments in USA.
reply
0 sats \ 1 reply \ @kr OP 26 Jan
Amazon lost money for decades investing/researching/iterating on automation and it's not like they're publishing the blueprints for their competition to read. Walmart has only just started investing more in automation after it became clear that Amazon's bet was starting to pay off.
do you think these kinds of fully automated warehouses have the same economies of scale that Amazon and Walmart’s human powered warehouses had?
or is there an outcome where robot-powered warehouses level the playing field between large and small merchants/suppliers?
reply
In USA, employees are extra expensive. Employers have to pay minimum wage, income tax, health insurance/benefits, safety accommodations, deal with the potential to be sued for injuries or discrimination.
Even non US employees will get sick, show up late, screw up at their job, steal things, have to be trained, they cause drama which impacts morale and they only work 8-12 hour shifts. Shopping for a employee is a process often requiring background checks, interviews, drug tests, contracts, and on-boarding rituals.
There will be a point at which automated labor out competes manual labor at everything and having human workers will be as barbaric and inefficient as having rows of women using typewriters to publish a report instead of using a inkjet printer.
At the same time, every household or community will probably have access to a local micro-factory (3d printer farm/factory) of its own so they won't need the services of the large mega-corp warehouses as much. Kinda like how the adoption of the microwave likely reduced the demand for restaurants.
reply
135 sats \ 0 replies \ @OT 26 Jan
There's at least one competitor in China called Jidong that also does fully automated warehouses.
reply
They would be fantastic bank tellers, but I think that's a while off. It's not that humans are better at it, but humans will prefer another human in a position like that which entails access to their money.
If I had an army of these robots, I might give dictatoring a shot.
reply
interesting, do you think people would prefer a robot or a human to serve them food at a restaurant?
reply