Kelly’s What Technology Wants


Jason Collins


September 19, 2012

Technology wants increasing efficiency, opportunity, emergence, complexity, diversity, specialisation, ubiquity, freedom, mutualism, beauty, sentience, structure and evolvability. As Kevin Kelly argues in What Technology Wants, these are the same things that life wants. Technology extends evolution’s four billion year path.

Whether you buy Kelly’s central thesis or not (in general, I don’t) and if you ignore some of Kelly’s near-religious fervour (particularly in the opening and closing chapters), Kelly provides a strong argument that the growth in technology is primarily beneficial, with the major benefit being increased choice. Technology provides the basis for achievement. The technology of vibrating strings provided the opportunity for virtuoso violin players. The technology of film allowed cinematic talents to blossom. And consider the technologies of writing and mathematics.

Kelly is not blind to the potential negative effects of technology. Shipping technology allowed mass slavery. The chemical industry spawns toxins. All technologies have unintended consequences. But in response to those negative elements, Kelly argues that prohibition is pointless. Prohibitions don’t work, don’t last and when they are put in place, are usually gone in the next technological cycle. Rather, we should use new technologies to offer solutions to old technologies, and use our knowledge of the path of technology to understand how to control the negative consequences. When new technologies emerge, they should not be banned but rather tested and actively assessed. Where harms occur, they should be rectified and where problems emerge, the technology should be redirected. The path of technology is inevitable and we cannot stop it.

The inevitability of technology is a central plank of Kelly’s argument. If the clock of time was rewound and started again, even with  different initial conditions, we would still end up with a similar path of development and resulting inventions. Kelly points to the examples of similar inventions occurring independently on different continents, such as agriculture. He points to the more recent phenomenon of multiple inventors of the same invention, such as lightbulbs or calculus. These technologies are inevitable, as is the rough order that they appear, as one builds on the other.

Kelly builds on this argument of inevitability by pointing to the (widely disputed) inevitability of biological evolution. Convergent evolution is a similar phenomenon to concurrent invention. Eyes and lactose tolerance evolved on multiple occasions.

I am not convinced that Kelly’s examples of concurrent invention or convergent evolution provide a strong case of the inevitability of invention or evolution, primarily because we don’t know what the fitness landscape of these technologies or traits looks like. If there is a single, clear peak for fitness, all paths will converge to it. If there is a complex multi-modal fitness landscape with a complex topography, we won’t see many of the possibilities. Within our own little world we will see convergence to a local peak, giving the impression of inevitability, but we might be missing the big picture. There may be an array of possibilities that we cannot get to.

Another issue is that we can find examples of one-off inventions or evolutionary solutions. As pointed out in a review by Jerry Coyne, the wheel only appeared in North America when brought by Eurasians. Similarly, bones, feathers and the human brain have only appeared once. How different would Africa or Australia’s path have been, and for how long, if they had been isolated from industrialising Europe?

I lean towards the view that biology is not repeatable . Small chance events have large effects. Although I am open to the idea that intelligence might be likely to evolve, a one-off example in over 4 billion years is hardly a strong case and doesn’t provide very many sample points.

If you pull that theoretical pin out of What Technology Wants, the argument that technology has direction lacks a solid thesis. As Matt Ridley did in The Rational Optimist, Kelly takes a general direction and tries to use evolution to turn it into an iron law. But it is the wrong tool to do so.

Regardless, I enjoyed the book greatly. It is full of interesting observations and ideas by an astute observer of technology. Just don’t look to it for the all encompassing theory of technology.