I loved his site when it was mostly about analyzing tech business models through the lens of Apple and his Aggregation theory. However, as his profile has risen, and his access to the tech elite increased (along with, possibly, his access to the ketamine they all seem to take), the quality of his analysis has declined greatly. Basically, he can reliably be expected to have the worst possible take on AI policy. We can speculate as to why (being an idiot, being smitten with tech CEOS, having large investments in tech stocks, or a brain tumor), but in the end it does not matter. Ben Thompson simply has nothing useful to say anymore about the tech market, unless you are looking to hear what the CEOs want you to think, but put more eloquently. Ironically, this is exactly the kind of the job that AI is tailor made to destroy.
As for his latest article, that robots can make everything does not mean that the owners of the robots will share the output of their robots. Not for free, or even at a price the poor jobless masses on UBI can afford. After all, becoming a member of the future 'robot owner class' is likely to be a byproduct of being a self-centered, win-at-all-cost asshole who values money and not people (Zuck, Besos, etc.) or at the very least someone willing to throw them under the buss to keep the money pump still churning (Cook if you are being generous). They will be either a Solipsist (who does not really believe other people are "real" in any sense that matters, and so is ok with their exploitation like Musk) or a narcissist (who admits they are real, but simply does not give a fuck - Zuck) willing to accumulate the necessary wealth to join that class by pressing as many people as possible to extract the value from their labor and make it your own.
The kind of altruism necessary to make his statement about ownership of the robots being a moot point true is antithetical to the accumulation of the power necessary to be a robot owner in all cases except one very narrow case... Government. If the government owned all of the robots (and was sane, had the support and interests of the people at heart, and was not co-opted by the very folks most likely to be the robot owners) it might be that ownership could be considered secondary to the output of the robots. But as long as those who own the robots are humans, who value money, social status, assets, or anything else that they might be able to extract from non-robot-owning humans, the question of 'who owns the robots" is a critical question.
Fact is, the AI revolution envisioned by Tech CEOs (and Ben) is going to destroy the ability of most people to make a living with their labor. And UBI does not solve that, it simply locks in the advantages for the owner class, and the disadvantages for everyone else. Think more "The Expanse" and less "Star Trek". The majority of humans living on below subsistence handouts from the ruling elites, except that the ruling class will not be answerable to those masses. They will only be answerable to themselves. Serfdom 2.0, except even worse for the serfs, because competence at a skill will be worth precisely nothing .No ability to pull yourself up or socially advance, because the table stakes will be to have inherited a double digit share of the global wealth.