Summary of "What Technology Wants"

Core Idea

Kelly argues that technology is not a pile of gadgets but a living system he calls the technium: tools, machines, software, laws, institutions, and culture that collectively generate more technology.
His central claim is that the technium has its own direction and partial autonomy; it is created by humans but not fully controlled by human intention.
Technology’s “wants” are not conscious desires but tendencies toward self-assembly, self-perpetuation, complexity, and expansion of possibilities.

Technology as Evolutionary Continuation

Kelly treats technology as evolution accelerated, extending the same self-organizing process that produced life.
He argues that evolution is not just random drift but also shaped by inevitability: physics, geometry, chemistry, and self-organization repeatedly funnel life toward similar solutions.
Convergent evolution is his key evidence: eyes, wings, echolocation, warm-bloodedness, and streamlined aquatic bodies evolved more than once in unrelated lineages.
He uses repeated features like rhodopsin, bilateral symmetry, and metabolic scaling laws to suggest that nature repeatedly “finds” certain structures because they work.
On this view, DNA is a uniquely powerful molecule for self-replication and complexity generation, and the technium is its technological extension.
Language is the bridge between biology and technology: it enabled cumulative innovation, making culture and tools evolve much faster than genes alone.

The Long Arc of Progress and Its Costs

Kelly insists that progress is real, though uneven and non-utopian: over long periods, societies get richer, longer-lived, more educated, and more morally expansive.
He links progress to expanding minds, cumulative knowledge, urbanization, science, and cheap energy, especially coal and oil, but rejects energy alone as an explanation.
He argues that science was the decisive accelerator because it created a better system for learning, storing, and distributing reliable knowledge.
Progress also expands moral concern, with the circle of “us” widening from family and tribe toward nation and beyond.
Cities are a major engine of progress: they are the largest technology humans build, and people move to them for freedom, opportunity, and specialization despite slums and hardship.
The costs of the technium are real: technogenic problems include pollution, climate change, toxins, obesity, species loss, propaganda, and nuclear risk.
Kelly says many technologies first create problems and then new technologies are needed to address those problems, in Brian Arthur’s sense that “problems are the answers to solutions.”
He argues that modern systems also erode direct human competence by offloading memory, labor, and even parts of daily life into machines.

How to Judge, Steer, and Live with Technology

Kelly rejects both naive celebration and total rejection; he wants people to coax technology toward its “inherent good” of increasing choices, freedom, and possibility.
He proposes reading technologies through 13 exotropic trends: efficiency, opportunity, emergence, complexity, diversity, specialization, ubiquity, freedom, mutualism, beauty, sentience, structure, and evolvability.
A technology is more “inevitable” and convivial, in his sense, when it advances more of those trends.
He uses mechanized agriculture as an example of something partly inevitable because it increases efficiency and specialization, while criticizing industrial farming for monoculture, fossil dependence, and ecological simplification.
His ideal alternative is not a return to the past but advanced gardening: decentralized, hyperlocal, highly specialized food systems layered with better tools.
Kelly’s preferred policy stance is a Proactionary Principle: anticipate, test, monitor, correct, and redirect technologies rather than trying to ban them in advance.
He argues the Precautionary Principle fails because the real effects of new tools are often unknowable until they are widely used.
He uses the Amish as a model of selective adoption: they are not anti-technology, but community-governed, empirical, and willing to use a tool in one context but not another.
Their rule of thumb is effectively “try first and relinquish later, if need be”, which Kelly contrasts with ideological all-or-nothing stances.
He also cites the minimite idea: use the least technology needed to do the job, though he notes that such simplicity depends on the larger technium outside the community.
Kelly takes the Unabomber seriously as a diagnosis of technology’s autonomy and self-serving momentum, while rejecting his conclusion that technology must be destroyed.

What To Take Away

The book’s big claim is that technology is a self-propelling evolutionary system, not just a set of tools under human control.
Kelly’s deepest optimism is that the technium tends toward more structure, diversity, intelligence, and choice, even when it also produces new harms.
His deepest warning is that technological gains routinely bring unintended, technogenic costs that must be watched, measured, and corrected.
The practical stance he favors is neither surrender nor rejection, but disciplined participation: understand technology’s direction, then steer it toward what enlarges life.