I don't see this one happening without major advancements in battery tech.
This is possible. Production of them will be the main limiting factor. I could see a situation where general-purpose robots exist for most tasks, that can do most worker's jobs better than the worker can...but it doesn't happen for a while because there simply aren't enough robots and/or they are too expensive.
Moreover, why would a humanoid robot be the most efficient robot? We aren't the most efficient design, we are a product of evolution. Tech doesn't have those kinds of limitations. I think it is much more likely that we have more and more AI enabled/controlled automation in our homes and lives (smart homes on steroids).
I think people underestimate the power of the humanoid shape for general-purpose uses. Yes, we are the product of evolution...but that means our shape is good at interacting with the world. Furthermore, our appliances and tools and buildings have all "evolved" to accommodate human-shaped beings, so at least at first, I think that's going to be the preferred shape for general-purpose tools. It's fine to have a Roomba-shaped thing if you just need a vacuum cleaner, floor scrubber, or lawn mower...but if you want a single machine that can do all those things and also load the dishwasher, do the laundry, fix the plumbing, cook dinner, and re-shingle the roof, then I think a humanoid will be the best choice. At least for a while, until our appliances/tools/buildings start coming in other varieties to accommodate other shapes of robots.
China will get there long before we will.
It's possible. I'm very worried that Trump's tariffs have made that more likely, especially if he doesn't back down soon. However, China seems to be having some internal political problems of their own at the moment, and it's unclear how serious they are or whether they'll threaten the country's economic trajectory.
One could also argue that if we were to grant more of our state and local governing functions like permitting and infrastructure design to AIs in the future, they would determine in most cases mass transit (streetcars, lightrail, autonomous buses) are far more efficient than hundreds of millions of autonomous vehicles.
I agree, for cities. For people who live farther out, I'm not sure mass transit will ever have the appeal. But since the driver's wage is one of the main expenses and that will be disappearing for autonomous vehicles, it's possible that I'm wrong. Maybe mass transit could become viable even in American suburbs and exurbs, once the cost of paying a driver goes to zero. I'm unsure. I suspect people in those areas would still rather use an autonomous vehicle for a while.
I think this one is just Elon's typical utopian nonsense. We can't change human psychology in 10 years, and markets are at least partially instinctual for humans - so everything won't be free.
I agree that markets will still exist, but I think what he's getting at is that most goods and services will become so abundant that the cost of producing them drops to near-zero, and therefore the cost of buying them drops precipitously too. "Near free" might be an exaggeration in that time frame though. Again, this probably goes back to your point about whether robot production is limited by other factors (e.g. battery tech). If there are a sufficient number of robots available to do a task and it isn't particularly energy-intensive, it seems the sky's the limit on how productive they can be.
