I think this is a serious, and probably quite elitist, misinterpretation of the reality and of human nature in general. As much as he's acting like a monumental fucking idiot right now, Musk is absolutely correct in his estimation that curiosity is the single most important human intellectual trait (and the thing AI safety should most directly address). Everybody is curious and/or passionate about something. And while boredom and depression can definitely be killers, these often arise not from simply having too much free time, but from constantly enduring a combination of financial strain and idle time. It's the combination of the two that is debilitating. That's why people so often conclude that welfare creates lazy freeloaders. The real problem is that welfare keeps people locked in a state of financial need because it never amounts to a living wage, deliberately imposes a feeling of guilt and shame, and leaves people without the resources to pursue their interests—unless those interests require zero financial support. With no spare resources to invest in the materials, services, transportation, education, and so on, required to pursue an interest, only those whose interests require virtually nothing (e.g., writing, some forms of art, sports, etc.) are able to find fulfilment.

"...people could work just enough to be fulfilled..."

Do you really think most people are fulfilled by the work they do to survive? I certainly don't. I think that's actually the biggest problem in our society today. We've created a culture of what is essentially indentured servitude—a class of human robots—with a huge percentage of the population doing jobs they hate while still earning barely enough to get by. That's the worst of all worlds.

I thoroughly believe that everybody has something that drives them. And if they don't now, given a little time and freedom from the pressure of financial survival, they would find it. I've noticed many times myself that when financial pressure is lifted I suddenly feel greater motivation, have more ideas, and see a world of possibility where I used to see barrier after barrier. Ninety-nine percent of the time, the problem is not whether a person has interests, but how to connect those interests to creating a viable income. That's where people struggle. Someone doesn't get into carpentry because they need a new table. They get into it because they saw someone doing it once and thought it looked super cool, or they remember enjoying it in school, or they have cool ideas for furniture, and so on. But once they give it a try—which, in itself, usually has a price tag—and "catch the bug", they often invest a significant amount of cash and resources into learning the skills, buying equipment, building or renting a shop, and so on. And too often, when considering the financial investment and the long road ahead, those who can't see themselves making a living from it, talk themselves out of even getting started. And that's when boredom and depression set in.

Now, I'm not necessarily saying that UBI is the solution, by any means. There is a certain satisfaction to earning your way in the world, and jobs taken for income certainly can become a source of personal self-worth. But I don't think they're at all essential for that function, and certainly don't think people would go to waste without them. Would anyone suddenly feel lost inside because they couldn't run around a warehouse for eight hours a day, seven days a week, working their asses off and pissing in a bottle, for a wage that barely pays the rent? Honestly, at the rate we're going, I think AI taxation and UBI are going to be essential. I just don't think they're the only tool required, and they certainly don't replace the need for extensive AI and robotics regulations (as well as much more serious commitments to breaking up tech/business monopolies and curbing runaway profiteering).

PS - I should mention that our capacity to achieve even Stage B is debatable. LLM-based AI doesn't reason, and that's been proven pretty convincingly, by more than one study. It more or less memorizes and has an extremely refined process for matching memorization to application based on context. Personally, I'm with those who think that a fundamental architectural redesign is going to be necessary to reach true human-level reasoning.