Jerry Kaplan

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI asks what it would mean to see the machine clearly, without the narcotic of hype or the paralysis of fear. Kaplan is the cycle’s economist of the artificial: the voice that grounds every conversation about consciousness and capability in the unglamorous question of distribution. His framework cuts through the two fantasies that dominate the discourse. The utopian fantasy holds that the rising tide will lift all boats; the apocalyptic fantasy holds that superintelligence will end human relevance. Kaplan dismisses both, not because he is indifferent to the technology’s power but because he understands it too well. The machines will do what they are built to do, and the question that should keep us awake is not whether they will develop a will of their own but what happens to human beings when the economy no longer needs most of them to function.

His lens reframes the orange pill moment itself. The twenty-fold productivity multiplier that engineers measured in Trivandrum is, in Kaplan’s analysis, a surplus that will flow somewhere—and the direction of flow is not technologically determined. It is institutionally determined, which means it is politically chosen, which means it can be changed. This is simultaneously the most sobering and the most hopeful implication of his work: the outcome is not written in the technology but in the rules we write around it, and we have not yet written them.

Kaplan stands in the cycle’s gallery as the thinker who insists on the economic floor beneath every philosophical question. Where Joanna Macy asks how we emotionally navigate the transition, and Joel Mokyr asks what institutional infrastructure will channel the gains, Kaplan asks the prior question: to whom do the gains currently flow, and how do we change that? His answer is structural rather than sentimental, and it is more demanding than either the triumphalists or the elegists produce, because it denies the consolation of helplessness and insists that the choice of which future we inhabit is still ours to make.

Origin

Born in 1952 and shaped by Chicago’s tradition of taking ideas seriously across disciplinary lines, Kaplan arrived at Penn’s computer science program already equipped with something rare in the field: a historian’s skepticism toward the stories a discipline tells about itself. He entered AI during the expert-systems era, when serious researchers believed that encoding the rules a professional follows would yield a machine that reasoned like one. He built those systems at Teknowledge, watched the first wave of optimism crest and collapse when the world turned out to be too messy for hand-written rules, and took from the experience a permanent suspicion of the field’s periodic certainties. Each generation, he observed, believes this time is different; each generation’s confidence is followed by what he taught his Stanford students as the pattern of AI winters—the predictable retreat when the world declines to cooperate with the lab’s best results.

His post-Teknowledge career was a tour of Silicon Valley’s defining moments: principal technologist at Lotus alongside Mitch Kapor, founder of GO Corporation to build the pen-based PenPoint operating system (a vision so far ahead of its hardware that the company became a famous and instructive failure), and co-founder of the online auction platform Onsale before eBay existed. Each venture gave him what the field needed and rarely got: the specific scar tissue of having built things that mattered, shipped products that found and lost their markets, and learned the difference between laboratory performance and durable value. His memoir of GO, Startup: A Silicon Valley Adventure, became a business classic precisely because it was honest enough about its own author’s defeat. A man who has written that book cannot be seduced by anyone else’s triumph narrative.

When Kaplan returned to Stanford as a fellow at its Center for Legal Informatics and began teaching the history and philosophy of AI, he brought all of this with him. The resulting books—Humans Need Not Apply (2015), Artificial Intelligence: What Everyone Needs to Know, and his later volume on generative AI—are not the work of an outsider alarmed by what he sees. They are the work of someone who knows exactly how these systems are made, who has signed the paychecks, and who is therefore in a position to say, with unusual authority, that the thing we should fear is not the machine but ourselves.

Key Ideas

Synthetic intellects and forged laborers. Kaplan’s most durable contribution is a taxonomy that cuts through the confusion of AI discourse. Synthetic intellects—disembodied pattern-finders that price insurance, screen résumés, recommend content, and trade securities—threaten white-collar work at software speed. Forged laborers—physical machines with sensors and actuators—threaten blue-collar work on the slower timeline that the stubbornness of physics imposes. The taxonomy matters because the two classes arrive on different schedules, demand different responses, and are most dangerously conflated. By organizing the field around capability rather than consciousness, Kaplan sidestepped the philosophical quicksand that swallows most AI debates and arrived directly at the question that actually matters for human lives: what happens to the people who used to do the work?

The ownership problem. Machines are capital. When a task that was performed by a worker earning a wage is performed instead by a machine earning nothing, the value does not vanish—it flows to whoever owns the machine. This is the engine of inequality that Kaplan placed at the center of his analysis. The standard hopeful story—automation has always destroyed jobs and always created new ones—does not guarantee that the new jobs will be as numerous, as well-compensated, or as broadly accessible as the ones they replace. And even when new jobs appear, the aggregate gains from automation accrue overwhelmingly to those who own the automating technology. Kaplan was unusual in naming this not as a malfunction but as the predictable output of a system working exactly as designed.

The robot uprising is a distraction. The apocalyptic scenario in which machines become conscious, develop desires, and turn on their makers absorbs attention and worry that ought to be directed elsewhere. Kaplan dismissed it not with impatience but with the precision of a builder. Machines perform tasks without wanting to perform them; scaling up the task does not spontaneously generate volition. The 2010 Flash Crash was caused not by malevolent algorithms but by perfectly obedient ones interacting in ways their designers had not foreseen. The danger is not rebellion but obedience—not misaligned consciousness but the absence of consciousness, machines following their instructions off a cliff because no one had told them where the cliff was. The billionaire who gravely warns of existential risk in the far future is conveniently not being asked about the warehouse workers his automated fulfillment centers are displacing today.

Job mortgages and ownership reform. Kaplan refused to leave his diagnosis at the level of complaint. His first concrete proposal—the job mortgage—finances retraining for displaced workers through instruments in which repayment scales with future earnings, aligning the incentives of lenders, employers, and workers in a way that existing financial infrastructure cannot. His second proposal reforms corporate taxation to make tax rates progressive with respect to ownership concentration: companies whose shares are widely held would pay lower rates, giving the entire economy a structural incentive to broaden rather than concentrate the ownership of productive capital. Both proposals are explicitly market mechanisms rather than redistribution—attempts to redesign the rules so the market’s natural operation produces a more broadly shared result.

The deepest question. Beneath the economics lies a question Kaplan approaches obliquely but never lets go: what are human beings for in a world where machines can do the work? Work has been the primary source of meaning, structure, and self-respect for most people across most of history. A society that solves the economic problem of automation—that finds some way to distribute the abundance the machines create—still faces the harder problem of what people are to do with themselves when their labor is no longer needed. Kaplan gestures toward an answer that his framework cannot fully supply: if the machines free us from necessity, the question of what humans are for becomes a question of what we would choose to do if we were free, which is at once terrifying and exhilarating, and which is, almost certainly, the question the AI transition is actually asking of us.