Defining Progress

“Progress” is a language game. The word does different work for different people: everyone claims to want it, or to be advancing it in some capacity, but almost no one means the same thing by it. Most definitions function as moral placeholders, letting people and institutions gesture at progress without committing to a claim precise enough to be checked: a kind of nebulous virtue signaling dressed up as conviction. This essay starts from the premise, inspired by Patrick Collison and the progress studies literature (see References), that we need an actual science of progress: a discipline for defining what progress materially means, and for telling real progress apart from progress that is illusory, or achieved irresponsibly; and that this need is sharper now, in the AI era, than it has been before.

The immediate provocation is the current Cambrian explosion of intelligence: capability compounding at extraordinary speed, but in narrow, spiky domains. That is not automatically the same thing as progress. Venture capital’s working assumption is that raw capability growth converts into a materially better world more or less by default. This essay’s position is that it can, and likely will, but only under conditions that have to be argued for, not assumed.