I’m writing this article about “Singularity Syndrome” before artificial “intelligences” are incorporated increasingly into the economy; I concede that this is a foolhardy approach. Days before the Wright Brothers launched their first flight, the New York Times published an article about how mankind would not master flight for millions of years. Similarly, at the beginning of the Internet Age, many economists and pundits believed the internet to be a passing fad that would have no greater effect on the world than the fax machine did. I’m not going to argue here that generative word machines won’t have an effect on the world, but rather that it may leave mankind’s capacity to innovate and develop increasingly stunted. It was predicted that when artificial intelligences took off that they’d be aware and intelligent, but as we’ve seen so far, that has not been the case. Drawing from existing materials, artificial intelligences have just given mankind an increased ability to draw from existing materials. For this reason, in this article, I will refer to the Large Language Models that have emerged as “reference machines,” rather than as artificial intelligences. If they are intelligent, they have not demonstrated it, and if they are aware, it seems to be of only existing work done by humans.
While the economic impact of these referencing machines will be profound, challenging both low-level workers and career-track professionals alike with new disruptions, it will also render the traditional career paths that society depends on for its continued function obsolete. If a referencing machine can conjure up different things in a field based on the information it has available, then it renders the education of people in that field meaningless and financially damaging within the capitalist system. People will not take on student loans to learn easily-automated things like accounting or journalism, because the rungs that once existed for entry-level workers to climb up will no longer exist. Instead, companies will probably lean more on skeleton crews of aging industry veterans to continue to pilot the organizations, while referencing machines do the majority – or entirety – of legwork that was once left to younger employees.
In such an economic environment, where learning becomes – by the nature of material conditions and economic incentives – increasingly obsolete, how much can we expect to expand in terms of our industrial base and productivity? In a world where thinking itself becomes largely automated, how much more discoveries can we really expect to make? Just as investigative journalism died as the internet rose to prominence, with the concept of investing money in journalists to go undercover and investigate things becoming irrelevant as headlines were simply copied in seconds regardless (leaving there zero competitive advantage for companies that did engage in investigative journalism), how will other careers be stunted in such a way? When the legwork consists of referencing existing work, what value is there in adding onto a body of knowledge that has no intention of rewarding or recognizing you for your contributions? As reference machines have shown us so far, contributors do not matter in the equation beyond simply expanding the machines’ parameters.
While some may point to the Open-Source Software community, in which thousands of programmers work anonymously together to create free software alternatives to corporate solutions, and say that landed professionals may still bother to contribute to man’s body of knowledge, I don’t bank on the idea of the labor aristocracy’s version of noblesse oblige changing things very much or lasting more than one or two generations. When the workers on these eventual skeleton crews retire, they will take their ladders and careers with them. At that point, the question becomes: are we even replacing the workers that had knowledge about any of this? And the answer to that is: probably not. The educational programs that would have lead to knowledge in those fields would have dried up, while the majority of mankind is left fighting over what little remains in the employment market. At that point, will there even be any profit in pursuing improvements and innovations in an economy made perfectly efficient? I don’t think so, because as we’ve seen as industries have become monopolized and divided up by oligopolies over time, at a certain point, maintaining market shares, defending market boundaries, and ensuring the smooth continuation of operations takes precedence over all else in terms of corporations’ bottom lines. In such an economic environment, innovation becomes obsolete.
With that being said, will reference machines alter the economy in massive ways? Probably. Will reference machines usher in a technological singularity? I don’t think so. The idea behind “Singularity Syndrome” is that, captivated by the concept of disrupting the traditional economy, the movers and shakers of our economy will also disrupt the foundations and processes upon which our society was able to develop and thrive, as machines simply become more efficient at capturing and digesting what already exists. Putting it in a metaphor, a creature cannot survive off of redigesting its excrement and I don’t think societies can either. As the pool of knowledgeable workers dries up over time, we will increasingly be less able to invent or contribute anything new because the systems’ support structures and incentives for doing that will only decrease over time. We’ve already seen that corporations prefer what is sufficient and predictable over what is novel, and in many ways throughout our lifetimes, we’ve already seen declines in standards of professionalism and production quality across many industries as the economic emphasis is placed more on replication than innovation. With that being said, putting the economic factors aside, how will the emergence of reference machines alter us socially and psychologically?
Reference machines, while great at generating a convenient string of words, may simply begin to operate at an economy of scale that will streamline people and products in ways we haven’t conceived of yet. Just as roads and mapping softwares lock cars into designated routes, with little to no thought or knowledge required of the drivers of the vehicles that use those infrastructures, people might find themselves similarly locked into designated patterns of thinking and behaving by reference machines. Just as debates today are settled by the list of results pulled up by internet search engines, it is possible that reference machines will similarly constrain discussion and activities increasingly. While the answers may already exist for the debates that take place, is it healthy for people to abandon their debates because a search engine draws up a list of conclusions for them to cherry pick from? I like to think that the debate, whatever it may be about, is part of what drives a society forwards, and eliminating that dialogue, however it happens, is dangerous.
While some people may argue about the validity of subjects like climate change and the moon landing, it only stimulates scientists and nut jobs on both sides of the coin to research further, learn more, and find new arguments. When we look at how even definitions and languages evolve over time, referencing an English dictionary published over a century ago in support of your argument about the meaning of something is patently ridiculous, because those meanings have likely changed since that book was printed. The greatest scientific theories that we have at our disposal today emerged out of debates and arguments, and for a time, those theories were the minority positions to take and had less references to back them up compared to the more traditional notions about how the world worked previously. All in all, how would reference machines alter this dialectical process in which we continually argue and refine what we know?
The answer is that we have yet to see this play out, but based on our past mishandling of digital technologies, my hopes aren’t particularly high for a future in which man finds himself increasingly dependent on reference machines. Francis Fukuyama once said that with the end of the Soviet Union, we had reached the “end of history.” Time marched onwards nonetheless. Here, I’m asking if these referencing machines will simply spell the end of knowledgeable discussion. It may not be apparent at first, but I think it is a genuine possibility as these technologies are increasingly adopted by the population and implemented across the economy.
