Raw LLM Responses

Neil's got it right. AI/new tech has already been replacing both white and blue collar jobs through attrition by making the workforce more productive. So IMO AI will only continue to make jobs more productive. There will be no big change, just the same continuous improvement as always. likewise. Humanoids have always been useless as "workers" and it's easy to explain why AGI can't change that. We see how Amazon is replacing people and it's not with humanoids. Because robots ALREADY use Vision and AI and they don’t look like humans because humans are limited. (they have to climb ladders, hands too big for small things etc.) Automation has been replacing people for decades and it hasn't stopped! So if a humanoid could do it, it’s already automated! (1) Just look at history. (2) So NO, they WON'T become useful as technology and AI improves. In other words.(3)In other words.(4) Many seem to think factory jobs are easier than other jobs, they are not. A lousy housekeeper won’t cut it. _____________Need more proof?_____________ Look at it this way. If AGI could assemble widgets or do surgery or manicures, etc. it would need special tools not batteries, hands and legs. Optimus is not strong enough to dig a hole or help someone get up off the floor. Even if it had a functional brain it still couldn't work at McDonalds. Can it wash its hands? How long would it take it to change gloves, start a roll of tape etc.? A humanoid MUST do a job on par with a person because workplaces are made so people along with machines maximize asset utilization. That is why you CAN'T just put two humanoids in place of one person. This is why no one has EVER shown ANY humanoid doing any useful job that a person was paid to do. But AI could make it a fun toy or criminal. If you have a trade job, you’d realize it would be impossible or impractically slow to do it with robot hands. No amount of AGI Learning or training is going to make clumsy hands better with tools, scissors, small parts etc. Making hands equal to a person is impractical and unnecessary. Hot, heavy, dirty and dangerous jobs. There are 100s of millions of them around the world. And if a humanoid could actually do it, LOL, most of those jobs pay super low wages in places where there is no electricity to charge it anyway. Most of today's humanoids can't complete the tasks from the DARPA robot challenge 10 years ago !?! "Optimus will be the best surgeon" !? When AGI can do surgery, it will use existing surgical robots. A humanoid will not wield a scalpel. "it's a humanoid so it can work in a human environment" The humanoid form or its brain doesn’t matter. What matters is what it can do. Who will buy them? Certainly not factory or warehouse managers. They can't do trade jobs or hard labor. They will be lousy housekeepers, tour guides, popcorn servers, entertainers etc. ridiculed as toys and criminals will use them for nefarious purposes. At best it could be a checkout clerk that will restock the store at night if it ever really gets a brain. Credit only, robot hands can’t handle money well. But they would be really nice if a handicapped person could use it as a telepresence to go shopping, do simple chores, etc. Please make a version like this despite "empathy is weakness" A better use for AGI in factories would be to replace low precision sensors. Example. If an AI agent could watch a conveyor line and all the parts on it, it could control the motors, escapements, pushers, etc. and it could see where things are without the need for sensors. is the escapement in place, is the part? what type, orientation, how many, did it fall? You could eliminate a bunch of sensors that are costly to maintain, it would save a ton of money on parts, engineering, programming, maintenance and downtime. It could be a natural language (instead of ladder or state) programmable logic controller. The new PLC will be NLPLC with the usual outputs, and mostly vision inputs. Keyence and Cognex already use vision to inspect, measure, identify etc. AGI could take it to the next level.(maybe they're working on it) you'd need several cameras at different angles for reliability/redundancy so AI is needed to identify targets from different angles and if cameras move. Maybe language programming would be a separate AI. It sounds like it might be a low bar for AI but I doubt it's really that easy. (1) It's first principles. If the goal is to emulate a human then you must make a humanoid/android. But if the goal is doing things that a human can do, then first principles DOES NOT dictate a humanoid. So AI is already "embodied" or "physical". (ill conceived terms, us old folks say, smart sensor or smart tv) And if you do want to emulate a human, you must choose which characteristics you want. It may need sensitive hands to feel slight imperfections or it may need durable hands for hard work. So a one-size-fits-all like Optimus is not practical. Unless it can reproduce, heal and recharge itself for decades by simply eating and sleeping.  (2) Warehouse managers needed people to store and retrieve items. Now warehouses are full of non-humanoid smart robots. People had to pick parts out of a pile and place them with correct orientation. Now one arm one eyed robots do it. People had to pick up parts to inspect and measure them. Now lasers measure them and vision systems inspect them. PID controllers became self tuning. As tech/learning/AI improved, so did machines and machine controllers. Just like computers taking jobs with spreadsheets and scanners. All the while someone was saying a humanoid will soon do the work. (3) Humanoids will never be able to do most jobs because most jobs are doing things that robots(software automation and hardware automation) cannot do and humanoids ARE robots. It's that simple but sometimes it's hard to see the forest from the trees. (4) AI has already been taking both white and blue collar jobs through attrition by making the workforce more productive. So IMO AGI will only continue to make jobs more productive. There will be no big change, just the same continuous improvement as always.