FAQ

FAQQuestions

Questions & answers

The Last Human Decision · Josh Luberisse Questions

Direct answer Straight answers, in the book’s own voice, to the questions readers most often ask about The Last Human Decision — what it argues, what it does not claim, and where to read more. Every answer on this page is drawn from the book’s text.

Will AI replace human judgment?

The Last Human Decision’s answer is: not “replace” — succeed. Machine intelligence does not need to rebel to inherit judgment; it only needs to become good enough that refusing to delegate becomes irrational. Institutions under competition could surrender consequential decisions one at a time, each step individually rational, while human authority remains formally intact. The concept the book coins for the endpoint is ceremonial sovereignty.

Does AI need to become conscious to take power?

No — and the book treats this as one of the most persistent sources of confusion. Institutions generally do not care whether a decision-maker resembles us; they care whether the decisions work. A market-making system need not feel greed to outperform a trader, nor a diagnostic model compassion to detect a tumor. Functional authority can arrive before philosophy settles what machines are.

Will there be a single AI takeover event?

Probably not. The book argues the transfer looks less like a coup and more like an invitation extended a thousand times: the system summarizes, then recommends, then acts and reports. Asking “at which point did the machine take over?” has no obvious answer — because nothing was taken. It was invited in at every stage.

What is ceremonial sovereignty?

The condition in which humans retain formal authority — titles, signatures, vetoes — while the consequential judgment behind those acts has migrated to machine systems. Like a constitutional monarch whose powers exist more vividly in law than in practice, the human sovereign persists as symbol and bearer of responsibility. The full explanation is here.

Why isn’t “human in the loop” enough?

Because presence is not control. A human can originate a decision, evaluate a recommendation, rubber-stamp it, monitor what already happened, or hold an emergency stop that is never used — all satisfy some ordinary meaning of involvement, and only some preserve judgment. A signature is not the same thing as a decision. See the five arrangements.

Is the book’s argument deterministic?

No. Its mechanism requires only three modest assumptions: machine judgment keeps improving somewhere consequential; delegation pays; competitors respond. The book is explicit that selection pressure is not fate — technologies disappoint, regulation forbids, societies preserve inefficiencies for the sake of values. The Selection Principle is described here.

Why would institutions delegate judgment at all?

Because competition prices the refusal. A company delegates because its competitor did; a military accelerates because its rival can act faster; a hospital adopts the system because the one across town did. Technologies can become compulsory without ever becoming mandatory: advantage becomes expectation, expectation becomes infrastructure.

Has judgment moved before?

Repeatedly. Craft workers conceived and performed the work; industrial management separated conception from execution; bureaucracy substituted rules for discretion; algorithms increasingly recommend what should happen. The book reads machine intelligence as the next step in that migration — not a rupture, but the boundary moving one floor higher. The full trajectory is here.

Will there be a machine ruler?

Maybe never — and the book’s more unsettling possibility is that no ruler is needed. Authority may disperse into procedures, workflows, objectives, and feedback cycles with no center to appeal to: rule by process rather than rule by someone. A machine monarch would at least be visible; process authority offers no throne to protest. The argument is here.

What is “the last human decision”?

The title’s answer: the final human decision may be to delegate one more decision. The transition the book describes requires no surrender and no dramatic moment — just thousands of individually rational choices by institutions that believe they are becoming more effective.

Who is Josh Luberisse?

Josh Luberisse is a technologist, entrepreneur, and writer whose work examines how emerging technologies reshape institutions, incentives, and human decision-making. He builds software and AI systems across a range of products, giving him a practitioner’s view of how automation moves from novelty to infrastructure. The Last Human Decision is his book about what happens when decision-making itself becomes a machine capability. More about the author.

Where can I read a sample of the book?

The Prologue (“The Succession Will Not Be Televised”) and the Introduction (“A Monopoly We Mistook for a Law of Nature”) are published in full on this site as free samples: read the Prologue or read the Introduction. The book is available for pre-order on Amazon — see the listing there for price, formats, and availability.