Elon Musk has said a version of this a few times now. In the future, nobody will need to work, because AI and robots will do it for us. Every time I hear it repeated I have the same reaction. That sentence is not really about technology. It is about ownership, and it skips past the one question that actually decides whether it comes true.
If you personally own the robots and the agents doing the work, then yes, you not needing to work yourself would be true. But that has always been true. It has nothing to do with AI specifically. Owning capital that produces value without your daily labour is exactly what owning a business, a rental property, or a factory full of machines has always meant for the people who own them. AI does not invent that arrangement. It just makes the machines smarter.
The question worth asking is not whether a machine can do the work. A lot of the time now, it clearly can. The question is who is legally allowed to own that machine, run it, be held responsible when it makes a mistake, and collect the value it produces. That is a rules question, not a capability question, and right now the rules are nowhere near ready.
What Argentina Actually Proposed
I keep coming back to something happening in Argentina, because it is the clearest real example I have found of a government actually trying to close that gap instead of just talking about it.
In May 2026, President Javier Milei published an op-ed announcing that Argentina wants to become the most attractive jurisdiction in the world for AI companies. Part of that plan is a draft General Companies Law that would create a new kind of legal entity, something being called a non-human corporation. Under the proposal, a company could be run entirely by AI agents, and human shareholders would be allowed but not required.
That is a genuinely bigger step than anything I have seen another country attempt. It is also, if you read past the headline, more limited than it sounds. Later reporting on the actual bill made clear that even in this deliberately deregulated proposal, a human administrator is still required to supervise what the AI run company does. The automation can run the company. A person still has to be answerable for it.
The Gap That Actually Matters
That detail is the whole argument, as far as I am concerned. Argentina is the government most willing, right now, to bend its own rules toward Musk's version of the future. It is actively trying to let AI own and operate a business with no person required to hold shares. And even there, the draft still keeps a human in the loop to answer for what the AI does.
If the most aggressively deregulated proposal on the table today still needs a person accountable at the end of it, then the distance between now and nobody needing to work is not a model problem. It is not a compute problem. It is a legal personhood and liability problem, and nobody has actually solved it yet. Argentina has only proposed edging toward it.
I am not saying Musk is wrong forever. I am saying the timeline being sold alongside that claim assumes the legal side moves as fast as the technical side, and everything I can see says it does not. Someone still has to be able to sign a contract, hold a bank account, be sued when something goes wrong, and be held responsible when an agent working under a company's name does something nobody intended. Until a jurisdiction actually lets a non-human entity carry that weight on its own, with no person standing behind it, the machines can be as capable as they want. The work of being accountable for them still belongs to a person, and that is still, in the most literal sense, work.