Exponential growth never lasts

The argument

The argument turns on a general observation about growth curves rather than on any specific forecast about artificial intelligence. In technology, in business and in society, growth has a tendency to end. Rates that look explosive when measured over a short window flatten as the conditions that sustained them are used up, and proponents of this position hold that there is no reason to treat automation as the single exception to that pattern. The mechanism they point to is the limiting factor: the variable within any field that prohibits growth once it binds. A limiting factor need not be a failure of ingenuity. Technology may remain perfectly capable of developing exponentially in principle and still be cut short in practice, because the curve depends on inputs that do not scale alongside it. On this view the relevant limits for automation are of three kinds. There are resource shortages, where the materials that machines are built from become scarce or costly. There are energy shortages, where the power required to build and run systems outpaces what can be supplied. And there are systemic constraints, of which climate change is the example offered — a disruption to the wider environment in which any economy operates, and therefore a brake that no amount of technical progress inside the field can lift. The conclusion supporters draw is conditional rather than triumphant. If automation's advance is bounded by factors external to it, then the displacement of human labour is bounded too, because the extrapolation from present capability to total substitution assumes precisely the uninterrupted exponential that never in fact persists. Because exponential growth never lasts, they argue, artificial intelligence will not absorb all work, and there will always be jobs for humans.

Context

While automation technology is currently growing faster than any previous technology, exponential growth never lasts forever and automation is no different. The sharp growth of it is a sign of the success of our system, not its failings.

Premises

Counter-arguments

The creation of an artificial intelligence unit that has 'general intelligence' (or about the equivalent of human intelligence) would quite likely overcome this simply due to its ability to harness more energy at a much quicker rate than we can. Our energy is limited, but in a crisis situation we could give an AI with general intelligence a million times human brain power and allow it to develop its own solutions to our resource limits.

Rejecting the premises

P1 - This growth period will not end soon because it is a unique technological field deeply entwined with the success of our society P2 - AI itself will help us to overcome limiting factors through exponentially self-developing