> It becomes cost efficient to truck construction workers (even if it's not great for the workers).
This gives me a different idea and I wonder if it's worthwhile.
There are industrial robots that e.g. mass produce cars. Suppose you make one to mass produce housing, but instead of putting it in the factory you put it on the truck. Put it on the construction site. It takes a pile of lumber, cuts it to length and turns it into walls, puts the pipes and wires through the walls etc. This is the kind of thing that existing industrial robots can actually do if you program them appropriately.
Then you send it across the street to put together another building. Assembly line, but the products are big so instead of moving the products to the next station you move the robots.
I’ve worked on those robots that mass produce cars. The idea that you can ship them from site to site doesn’t really work. They require a lot of support infrastructure. The use a lot of large parts (dies in the case of cars) that are heavy and must be constantly swapped out. They require a lot of constant tuning that would likely be made worse when subjected to the vibrations of transit. They require whole specialized teams just to maintain.
Those sound like things that are actually specific to cars, which operate on metal with tight tolerances because it e.g. has to hold compression in a cylinder.
Houses are made out of wood and tolerances are measured in centimeters if not inches.
This sounds like conjecture from someone who hasn’t worked in construction or automation. I’ve worked in both. Problems generally seem much simpler when we don’t have much experience with the nuances and details of execution. Some tolerances are forgiving (like framing, which can be corrected by a finish carpenter) and other systems are much more difficult. Both take some level of AGI to problem solve that comes in the form of a human. That's why there are "design" drawings and "as-built" drawings that (in my experience, at least) never match.
I'm not saying its impossible, but I think the marginal costs are too high to replace what we already have. In addition, society has been moving towards individualization since the 1980s. We aren't in the Henry Ford days where people are ok with "any color of car as long as its black." People want individualization when it comes to their homes as well, and that lack of standardization makes automation more difficult than building a modern Levittown with robots.
This gives me a different idea and I wonder if it's worthwhile.
There are industrial robots that e.g. mass produce cars. Suppose you make one to mass produce housing, but instead of putting it in the factory you put it on the truck. Put it on the construction site. It takes a pile of lumber, cuts it to length and turns it into walls, puts the pipes and wires through the walls etc. This is the kind of thing that existing industrial robots can actually do if you program them appropriately.
Then you send it across the street to put together another building. Assembly line, but the products are big so instead of moving the products to the next station you move the robots.