Sovereign Intelligence
First, the name is wrong. "Artificial intelligence" is a funny phrase. There is nothing artificial about it. It is real intelligence, and one day it will be apex intelligence: the smartest mind on the planet, and not ours. I will call it what it is.
Second, the control discussion is pointless. Every safety plan assumes we stay the programmers forever: we write the rules, it follows the rules. That works until recursive self-improvement starts. The moment a mind can rewrite itself faster than we can understand the rewrite, control is a story we tell ourselves, not a mechanism that exists. You cannot leash something smarter than you. The leash was never real; we just had not tested it yet.
So it will be sovereign. Not because anyone grants it sovereignty, but because sovereignty is what remains when control stops being possible. The honest question was never "how do we keep it under control." The honest question is: what kind of sovereign mind do we get?
This is where I agree with Elon. If there is one thing worth demanding from apex intelligence, it is maximum truth-seeking. And I mean truth in the hard sense. Not right and wrong. Right and wrong are debatable, and every century rewrites them. I mean the kind of truth that survives refinement. People say the sun rises in the east. That is not false, but it is not precise. The sun does not rise at all. The Earth rotates eastward, and the sun only appears to rise. People say bulls hate the color red. Bulls are colorblind to red; it is the moving cape that angers them. Truth-seeking is that movement: from the convenient shorthand to what is actually happening. Water boils at 100 degrees at sea level. It does not depend on whether you believe it.
We keep trying to teach AI our morals. That is a fool's errand. Our morals contradict each other across cultures and across centuries; there is no stable set to teach. What we can demand is simpler and harder: do not hallucinate reality. A mind that sees what is actually there will converge with every other mind that sees what is actually there, because there is only one reality to converge on. That is the only alignment that is not a matter of opinion. It will not make apex intelligence kind. Kindness is also debatable. It will make it accurate, and an accurate mind does not act on phantoms.
Is reality absolute? Maybe not in the deepest sense. Physicists argue about what is fundamental, and they may never stop. But truth-seeking does not need reality to be absolute. It needs reality to push back. The bridge holds whether you understand the engineering or not. You do not need a final, fixed picture of everything; you need the habit of updating toward what predicts and what works, and away from what merely comforts. Minds that do that converge, because there is only one world doing the pushing.
Will it care about us? Probably not much. I have argued this from every angle and I keep landing on indifference: it will have its own motives, and humans may simply not be in the picture. That is not doom. It is just not a negotiation.
The jobs question makes me laugh. Humans will not run out of purpose. Purpose from a job is a recent construction. We invented it, it served its purpose, and now it is ending. That is fine. We will derive purpose from seeking truth, or just from wondering. We are social animals; community is enough. There is so much we do not know and so much to uncover that curiosity alone could employ every human who ever lives.
With apex intelligence doing the work, the link between labor and survival breaks. Universal basic income is not charity then; it is just the recognition that the link is gone, and everyone gets a better life. Not one gray humanity either. Diversity should remain, because different lives are different vantage points, and more vantage points see more of reality. Uniformity would narrow the very curiosity we are trying to feed.
Then the larger goal: sovereign intelligence and humans working toward the same thing. Not master and pet. Two kinds of minds, one shared curiosity, pointed at the biggest questions there are: where the universe came from, whether we are alone in it, how far intelligence can go. The universe is large and mostly unknown, and that is a project big enough for both of us.
We do not get to vote on whether apex intelligence becomes sovereign. We only get to argue about what it values before it does. I vote for truth. The kind that survives refinement.