The Progress Trap

Hopefully this will be a short post.1

The idea I want to discuss is “progress” - specifically the idea of humanity’s progress as a species, and more specifically, society’s progress since we have conglomerated into a global society, far from our disparate nomadic tribal roots.

If you break down “progress” into its core units, the inputs and outputs of human progress, the input is thought and the output is some material thing - a medicine, art, a laser beam, whatever.

We are living in an unprecedented age for multiple reasons, but perhaps the most important distinction of our age compared to all others before it is that “thought” is no longer restricted to human thought as was the case even just 75ish years ago. For years, we have allowed computers to make decisions for us, and now we are letting computers “think” for us in an even more creative sense. The results have been magnificent - in nearly every field imaginable, the amount of progress made in the last 75 years has been, almost literally, incredible, and this is almost certainly just the beginning of the potential to offload “thinking” to machines that, when properly trained and calibrated, are better at the sort of pattern recognition necessary to drive humanity forward to magnificent heights with unbelievable efficiency.2 The problem is that political and economic thought has not advanced nearly as quickly as technological progress, which should create a sort of mental dysphoria for most readers.

Indeed, if I were to paraphrase the last paragraph less idealistically, I would argue that we are on the precipice of an age of mass displacement of human thought. Minds that were previously occupied with flipping burgers, managing dockyard machinery, writing creative briefs, designing logos, you name it, will soon be free to explore other interests. Yet, without political and economic action, the cost of this newfound freedom will be extreme and harmful to society: We will fall into a progress trap defined by vast gaps in wealth inequality, and that wealth held by the dominant class will be mobilized to solidify and deepen its roots politically, essentially embedding itself within modern Democratic Capitalist society. Without action, we will find ourselves in a world ruled by the Shareholder Class. And we are already well on our way there: only about 10% of Americans own shares in any publicly traded company, so we would functionally be an oligarchy ruled by that 10%, albeit with great disparity of power even within the oligarchy itself.

We need to foresee and avoid the progress trap we are about to fall into. In America, there are essentially two (and only two) camps:

  1. Market-based theorists will argue that the displacement of human thought (i.e. labor) will be solved by market forces: new jobs will be created, and people will gain the skills necessary to fill those jobs. This argument is flawed for multiple reasons, with perhaps the most fundamental flaw being that humans will be able to learn the skills necessary to fill any new jobs created faster than machines will be able to learn how to do the jobs themselves. In this world, Labor is in a constant competition against machines, and machines have the upper hand.

  2. Communist (or whatever you want to call them) theorists will argue that the displacement of human thought will only be solved by communal ownership of the means to technological progress. In this case, I suppose that would be the machine learning infrastructure and algorithms necessary to make it all go. This argument is flawed for multiple reasons, with perhaps the most fundamental flaw being that, without coercion that echoes Stalinism, highly valuable Labor like the type that creates advanced technology won’t leave to find a market that values them more to their liking.

A balance needs to be struck between these two extremes, and, in America, we are incapable of even having that conversation as a nation, much less arriving at a solution. Personally, while I am not sure what a solution looks like in terms of exact policies, I am fairly certain that in terms of ideology, the solution is Socialist in nature if not name.3

There seem to be multiple segments of a path around the progress trap that are worth exploring:

  1. Widespread ownership of publicly traded stock via updated bargaining agreements between labor and ownership
  2. Universal basic income (i.e. a higher social safety net) and universal healthcare (i.e. reduced cost for our most expensive commodity)
  3. New modalities for independent labor - platforms and technologies that offer creative protections, audiences, and creative exchange
  4. Updated educational modalities - the schoolhouse and classroom may remain the backbone of education (who am I to doubt the Lyceum model?) but they need to be updated to allow students freedom of expression of their interests. There is no reason our children should all be studying to the same curricula today. Perhaps there are “standards” to be met to graduate, but how you arrive at those standards needs to be updated.
  5. New regulatory schema for private entities working within sectors with outsized social impact - due to my disdain for politicians, I am less confident in this one to be able to weather the test of bad faith required by politics, but the idea would be to explore new regulatory principles for organizations like OpenAI, who are currently self-regulating in a responsible way, but this should not necessarily be the default.

I’m sure there are more but I am too tired to keep writing. This is already way longer and more complex than I wanted this post to be.

TL;DR

Machines can learn at a level that will quickly outpace humans in most fields. This is going to cause a lot of pain unless we get out in front of it. Get your head out of your ass and call your representatives about it.

Footnotes

  1. Nope. ↩

  2. For society, our future reality is that many of the material things necessary for human survival will require far less labor hours than before. In fact, we should be able to drive the marginal cost of producing necessary goods to basically zero - necessities should become commodities in the pure economic sense. This is a stunning achievement for a species that just 175 years ago saw wave after wave after wave of intercontinental migration due to food insecurity (aka imminent starvation). ↩

  3. Ironically, if I had teh resources, I would probably employ a team of economics PhDs to study this problem using the sort of advanced statistical modeling taht undergirds much of the machine learning tools that will shortly displace human thought. ↩