Macro Economic Thoughts On Humanoid Robots

I am well aware, which is why I mentioned “low-cost” for the tools used by the humanoid robot. When you start getting into production volumes or ultra-high precision work, then it becomes cost effective to have dedicated robotic systems such as CNC, etc. I was using (semi-?) automated (pre-CNC) systems in the 1982-84 time period because it was both cost AND time effective. The machine was located in Chicago (450 miles away), but that didn’t matter. In 2000-2001, the technology had really progressed. Cost savings were dramatic compared to just a few years previous.

1 Like

Walk the dog? :smiling_face:

What differentiates humans from most everything is adaptability. It would not make sense to make a single purpose humanoid robot. Simpler machines can do simple jobs much more economically. The only logical use of a humanoid robot that I can see is a Jack of all Trades.

Why a humanoid format? Elon said it clearly, because we designed our world for humans to use and inhabit. The more the robot is humanoid the better it fits into our world.

The Captain

Walk the dog? No! It’s also a robot! :lollipop: [LOL e pop]

2 Likes

Which is why I think this is a weird product.

Most manual labor/low-skilled jobs aren’t “Jack of all Trades” kind of jobs. You’re doing the same handful of things over and over again. You’re breaking down chickens in the poultry plant or working the fry machine or passing items through the cash register at the grocery store. You know - the “boring and repetitive” jobs that were discussed upthread.

It’s true we have designed our world for humans. But for the most part, that just affects the rough size that a robot needs to fit into. If you have an AI that can do all the things a human grocery cashier can do (for example), I don’t need you to spend the extra money to give that AI a chassis with legs and the ability to self-balance or sit on a chair. I’d rather just mount it on a pole bolted to the floor behind the register and give it more than two arms, so that it’s moving and scanning on the front end and bagging on the back end.

Or honestly, if we’re spending tens of thousands of dollars on a robot cashier in the first place, maybe just integrate those capabilities into the the feeder belt assembly, rather than a separate unit standing to the side of the belt and working a cash register. We can get rid of the UPC reader (since the AI can scan codes with its optics), and just mount arms onto the middle of the belt (to reposition items so the AI can see the UPC and lift/weigh things like produce) and at the end to bag items.

After all, the reason we designed that tiny part of our world (the checkout line) for humans to use and inhabit is because we need a human to scan and bag the items. If you aren’t going to have a human do that, then it’s worth redesigning it to fit the AI that will be doing that. It’s short-term and inefficient thinking to try to shoe-horn the AI form factor to fit the existing, human-optimized checkout line, rather than reconfiguring the checkout line from first principles.

I don’t know what came of it but a grocery store here tried just that (haven’t been there in awhile). The customer unloaded the cart on a long feeder belt with items separated, and it passed under a scanner and a camera and would automatically scan items, which feed another long feeder belt where the customer did the bagging. Honestly, other than the capacity to handle a large cart’s worth of food rather than “15 items or less” not sure why this was better than self-checkout.

From your question to a TMF topic?

Why limit the "picking, scanning, loading to the standard grocery store format?

Google reveals LOTS of robots being trialed.

The self check is a money saving, “grooming” stop gap.
Robots have been “coming for grocery, produce, etc” jobs since… forever?

IMO.
:alien:
ralph

1 Like

I strongly suspect that the task will conform to the ability of the AI and robots, instead of the other way around. That is why we have UPCs on all of our groceries nowadays, and we no longer spend hours and hours learning to push numbers on cash registers. I don’t see the point in humanoid robots–I expect robots will be designed to do a task, and humans will continue to be the generalists.

2 Likes