Raw LLM Responses

​@ohwormman Ok, so to use my local shopping center as example. You have 4 trolley bays, which is the largest section in the entire shopping center, the average is 2 and many sections barely have room for 1. ALDI - 1 cart type Coles - 3 cart types, bigs and seats dont fit in smalls Woolworths - 2 cart types, bigs have seat variant but is otherwise the same cart Kmart - 1 cart type Target - 1 cart type Misc brands - 3 carts, none fit in any other cart Pray tell, how do you plan to make this better for the AI? Any amount of sorting will require the AI to have room to pool trolleys. A worker will normally just throw them out onto the road and rapidly toss trollies into safe zones, then rapidly return them to the bay. Imagine a robot shape that leans into doing this. As someone who's been friends with a lot of people who've studied robotics in university. Every single joint you add with double the cost of the robot. So you want as utterly simple as possible because if you say, have a three joint robot, you are probably looking at some horrifying 150,000 dollar robot that breaks down every month in these conditions and cost 20,000 to fix. This is because its gotta also have strength to move multiple trollies aroumd. if they can only do one at a time, I'd estimate you'd need 10x as many minimum, so this would likely cost the entire income of the entire shopping center just for trolley pushers. If you remember those cute dog robots that boston dynamics have? CSIRO, the Australian governments big research organisation that does a lot of very front of market robotics products, has two of those dogs. They cost around 250,000 aud per year currently. I know this because I asked the head of engineering for this multi billion dollar company so I can assure you I am right. And that's for one robot that replaces one worker. You'll actually want 5 of these, as that replaces the current 10 (5 for morning 5 for night) If you make the trollies themselves able to move, then it becomes unprofitable. The current 3000 trollies at my local shopping center (medium size, nowhere near the size of American mega shopping centers or the stuff you see in Asia) cost around 2 million to purchase, and generally break down once every 2 years and break down permanently within 5-10 years. When I asked my old boss, he said each trolley, containing a wheel lock, costs 2400 AUD or around 1100 USD. Just adding the electronics for the wheel costs around 200 bucks and halves the lifetime usability of the trolley. It is almost always what breaks first. Those 3000 trollies thus cost around 2 million AUD and average around 650,000 per year thus in replacement costs thanks to one single piece of electronics that is extremely simple. Your looking at 10x that for AI self drivers. Remember, doing this, is possible. The goal isn't to do it, its to make it cheaper than the 200,000 ish per year for the 10 people per day that currently do the job.