After Productivity
On AGI, upskilling, and what remains for most of us
A familiar defense of AI says greater productivity is a public good. More output means more possibility. Tasks disappear and better tasks appear. The story mirrors earlier shifts. Farming gives way to factories. Factories give way to labs and studios. The ladder keeps extending. The assumption is that the next rung will be there again.
The pursuit of general systems complicates that story. If the goal is a technology that can perform the broad universe of cognitive tasks at a competent level, then the ladder does not simply extend. It bends. A model that matches or exceeds an average human across a wide set of activities does not only free people from one domain so they can climb to the next. It follows them up the ladder.
Earlier revolutions redirected effort. Tractors altered agriculture, but they did not abolish human planning. Spreadsheets automated arithmetic, but they did not remove judgment from finance. The promise of general systems is different. It is not a single tool for a single layer. It is a companion that travels with the work, from research to writing to design to decision support. That raises an uncomfortable question. If the next industry appears, what exactly are most people upskilling into when the same class of system is also the best candidate worker inside that new industry.
Defenses remain. Some tasks require bodies in the world. Some roles depend on trust, taste, and the particularity of human presence. There is likely a persistent top tier of work that asks for rare judgment and creates new frontiers. That tier does not vanish. Yet the core concern is about the rest. If a system delivers reliable competence at low cost, the rational assignment for most complex tasks is to the system, not to people. In that world, humans do not disappear. They become a term in a cost function and are selected when the price is right.
The opposite extreme draws a different picture. People will never be obsolete because the demand for human meaning expands as material needs are met. New markets form for care, experience, and attention. The argument is not trivial. There is evidence that people value the human source of certain goods even when a machine can produce them. But this answer often assumes what it needs to prove. It imagines that preference will continue to create enough work for most people, at wages that can support lives that feel dignified.
Perhaps the truth resides in a redistribution rather than a victory by either side. AI lowers the cost of competent labor in many domains. It raises the premium on a few things that do not scale easily. Agency paired with expertise remains scarce. Taste that organizes abundance remains scarce. Legitimacy that people grant to one another remains scarce. These scarcities become more valuable. They also concentrate in fewer hands.
If this is even partly right, then two tensions shape the next decade. The first is economic. A small number of people can produce and maintain systems that once required many. The total pie may grow, but the labor needed to bake it shrinks. The second is cultural. If technology is good enough at everything that looks like work, society must decide what counts as meaningful contribution. That decision is not purely technical. It is political and aesthetic. It involves institutions, norms, and shared stories about why a life matters.
There are still practical questions for the present. What skills compound when competent systems are everywhere. One answer is judgment about what to build and why. Another is the ability to steward systems in the world rather than only inside the sandbox. Constraints from people, law, physics, and ethics do not evaporate. The surface of work changes, but the friction where technology meets reality continues to demand careful minds. There is also the slow craft of taste. In fields flooded with adequate options, selection becomes creation. The person who can choose well can still create value that is not easily priced.
None of this resolves the contradiction. The productivity story remains attractive because it has been mostly true for two centuries. The generality story challenges it because the tool now follows us into the new domains we would have fled toward. It is possible that both are true at once. Output rises and new wonders appear, while the distribution of purpose tightens. It is possible that we find new forms of collective work that absorb human energy in ways that are not yet visible. It is possible that we do neither, and we spend a decade confusing motion for progress.
Open questions are unavoidable. If competent systems perform most work at low cost, how do we ensure that human lives retain a sense of earned meaning. If the number of high leverage roles shrinks, how do we widen the path into them without pretending that everyone will walk it. If productivity climbs while the felt need for human labor declines, what replaces the social function that work once served for most people. None of these questions obliges a single answer. They do ask us to admit that the old pattern may not repeat without help.
For now, a sober stance is simple. Welcome the lift in capability. Refuse to let the word productivity stand in for a theory of the good. Treat general systems as partners that expand what can be built. Treat society as a project that must decide why building still matters.
Disclaimer: These essays begin as voice recordings, later transcribed and shaped into written form by AI. They express only personal reflections. Nothing here constitutes legal, financial, or investment advice, nor do they represent the views of any institution or individual other than the author.