AI as an attempt to play the endgame of capitalism
Today I saw a video from Rick Beato, whom I follow for his interview with interesting musicians, and his takes at checking out Spotify’s top-ten streaming charts.
Whatever you think of the current incarnation of AI technology (generative models, diffusion models), they will be and are being used to do the jobs of humans, in other words to replace human labour. At this point we’re talking about (with varying success)
- Juniors in software development -
see for example this excellent take on Stackoverflow’s blog on possibilities and limitations. - Artists doing one-off works for advertising, illustration of articles and the like -
here it seems that AI generated work is still easily recognizable. - First-level support which has been under pressure for several years already and with the ability to connect large language models with semantic search capabilities might actually be pulled off to be fully automated
- and, as exemplified by the above video, music generation based on known genres.
I’ll be honest here, while Rick says his son is able to recognize AI work, I am really having difficulties distinguishing the work from existing chart entries. The fact that the popular music industry has worked diligently to make pop music as bland and interchangeable as possible surely helps in the process.
I am sure you can come up with other areas where generative AI is set to have an effect in the next years.
What has this to do with Capitalism?
Today’s incarnation of capitalism appears mostly driven by the Friedman doctrine. The one thing a business is meant to do is to increase its profits. If you follow this doctrine in all of its consequences, the removal of human labour as a limiting factor in terms of costs and general hassle (unions, weekends, vacation, sickness, …) is a logical conclusion.
This sole focus on only one factor of what actually makes a business successful in the long term leads to a myopic view onto many factors that are also relevant.
Removing the human factor from how profits are generated may yet be the most short-sighted action of them all. Here we see a new technology which appears to be put to use to do just that. Historically, jobs have been lost to technology. This current trend does seem different though.
Generative AI is a foundational, generic technology that has the potential to seep into a wide range of human labour activities. Even one of its perceived major weaknesses (the fact that generative AI can make up things that do not make sense under the scrutiny of facts and its ability to make vague statements) is in fact shared by us and hence does not feel to me nearly as dangerous in everyday work as some people make it out to be. Thinking this through, you are left to wonder about two questions. One is related to a profit-driven capitalism in general:
Who will buy all those consumer goods if nobody is earning money?
The other is related to the use of generative AI in particular:
Who will create all the things that generative AI requires to generate its outputs?
Generative AI only works because the underlying algorithms were fed with a sheer unimaginable corpus of training data containing facts, information, insights and art - artefacts that were created by humans. That corpus was not only generated by altruistic talents who want to benefit the whole of society, but also by people who gave their hard work to earn a living. If that was an important motivator for those people, what will motivate them to create content in the future?
These are externalities that cannot be ignored, yet the relentless pursuit of profit often leads businesses to do just that.
What has this to do with the Butlerian Jihad?
The Butlerian Jihad is a fictitious event in Frank Herbert’s Dune Universe that occurs about 10 millenia before the events depicted in the main novels (or the newest movies).
It is a war between humanity and thinking machines whose outcome was the ban of intelligent machines.
Though I’m typically a technophile—believing that developing tools is one of humanity’s greatest strengths, I can imagine that such events could unfold.
In Herbert’s work the machines were also in the business of ruling which apparently were one of the trigger factors for the uprising against them. Here I believe that certain people who have too much of a desire to rule would put a stop to governments based on machine-led decisions (then again, if profitability is the ultimate maxim, what are high-frequency trading algorithms in this world?).
However, the current incarnation of applying AI also carries in it the seed to reduce the available options for many humans in terms of what work is available and how they can participate in the current system and find their place in terms of purpose as well as simple economic wealth.
The Butlerian Jihad’s relevance today lies in its cautionary tale: as AI advances, we must be vigilant about the societal and ethical implications to avoid a future where technology controls humanity rather than serves it.
Given the right framing and unfolding economic conditions, a “Butlerian Jihad” may not be such an unrealistic proposition after all.