The complexity for minimum component costs has increased at a rate of roughly a factor of two per year. Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. – Gordon E. Moore, April 1965
Moore’s law was the gently hissing silicon heart of the tech sector for 5 decades. For 50 years we could expect large, consistent in the performance of our underlying hardware. The steady tick-tock of hardware development guided trends within the industry, and those trends are what made the culture and ideology of the industry. When hardware advances could be counted on to “fix” deficits in our software, it was easy to ignore the problems and push on to new features. When computing power had no meaningful upper bound, it was easy to assume that any problem would be solved eventually just by applying enough transistors to the job.
Neoliberalism, summarized very simply, is an ideological framework holding that state economic intervention is usually best avoided except to create markets, and that healthy markets serve to resolve many social issues via competition and rapid economic growth. It’s almost entirely an elite ideology; very few people outside the structures of political and economic power have neoliberal ideologies. Neoliberalism tends to be very concerned with quantifiable macroeconomic measures, rather than more direct measures of social outcomes or “fuzzier,” less quantitative measures.
One school of historical thought, generally – though not exclusively – associated with Marx, is called historical materialism. In short, historical materialism holds that history always arises specifically from the economic structure of society: who controls what, how the means of production are worked, how they produce things, and so on.
I’m not sure I’d call myself a Marxist, and strict historical materialism is a very imperfect tool. However, using a less dogmatic version of the idea as a lens through which to examine history can be useful, and that’s how I’d like to look at Moore’s law.
Tick, tock, tick, tock…
By the mid 90s at the very latest, the culture of the software industry was leaking out into the wider world in a serious way. By the 2000s, we were a serious cultural and political power. By the 2010s, we were the cultural and political power. As of writing, the top 5 largest companies on the planet by market cap are all tech companies..
But in many ways, we were in the driver’s seat much earlier. Computing was being used to great effect by governments as far back as the 40s. Long before it was part of mainstream culture, elites and technocrats were very interested in computing’s potential applications. And it’s here, I think, that Moore’s law gave such a strong tailwind to neoliberal ideology.
The attitude in tech, exemplified by people like Paul Graham, has been that many problems will just be fixed by the inevitable forward progress of computing. “We shouldn’t have to worry about efficiency,” they believe, “because computers should just get fast enough that it doesn’t matter.” This mirrors a key failure of neoliberalism, the failure that the financial crisis of 2008 laid bare: after a while, you have to keep growing just to stay in place. The moment growth stops, everything falls apart because you’d just used that ever-upward line on a graph to plaster over the rot.
The timing of the two, tech and neoliberalism, is striking. Contemporary neoliberalism really began in the early 70s. Intel was just a few years old. C and Unix were brand new. Computing as we understand it now was newly-escaped from the labs and just starting its life in industry. Neoliberalism and its center-left technocratic child were built around the productivity boom computerization caused across the economy, and around the new, sophisticated kinds of analysis high-performance (for their time) computers enabled.
Moore’s law, and more generally the culture of “up and to the right” in the tech industry that was swallowing up the rest of the economy, gave neoliberalism the solid ground it needed to go from political movement to dominant ideology of the global economic powers. The promise of perpetual growth was absolutely necessary to prop up neoliberal policies’ logic. As long as logistics was always going to get better, finance was always going to get better, manufacturing was always going to get better, many people trusted the reassurance from on high that “a rising tide lifts all boats.”
The logic played out everywhere, not just in macroeconomics. What passes for a business model over at 1455 Market St, what has somehow been enough to draw tens of billions of investor dollars, is literally just “we’ll stop losing money if we scale up enough.” It’s totally incoherent because Uber’s business isn’t subject to economies of scale at all, but because many investors have bedrock an ideological assumption that you can grow out of anything, they get away with it.
And today, Moore’s law is dead.
I don’t want to ascribe too much importance just to the details of CPU improvements, but it’s poetic that the most visible signs of neoliberal hegemony’s collapse, the Brexit vote and Trump’s election, came in the same year Intel finally admitted that its tick-tock cycle of processor development was dead. In computing, in resource extraction, in carbon output, and in so many other areas, neoliberalism has made like Wile E Coyote, looking down and discovering that he’s already run out off the end of the cliff and now there’s nothing but miles of empty air underneath him.
The perpetual growth ran out, and because too many of its adherents drank the “End of History” Kool-Aid, neoliberalism did not evolve to respond to the coming crises it had made. We in the technology sector bear more responsibility for that than I think most of us admit. Too many either drank the Kool-Aid too or, more cynically, realized we could profit from the faith in our ability to sustain permanent growth. That combination of willful ignorance and cynicism is where the Ubers and the Facebooks of the world came from.
Techno-utopianism is dying, finally, but we still have a tremendous task ahead of us. Bending software away from this pathological neoliberal capitalist model will mean far-reaching and painful changes in how we operate. That’s materialism, after all: ultimately, what we do is a product of how we do it. We need new models for the right way to build software, and a new ideology to reclaim technology for human benefit rather than capitalist scaffolding. I have no idea what those changes will even look like, but as Ursula Le Guin said,
We live in capitalism. Its power seems inescapable. So did the divine right of kings. Any human power can be resisted and changed by human beings.
We created this, and if we’re willing we can also change it.