The future of work, as I see it
To understand the future of work, it helps to think briefly about the history of work.
- in pre-historic times, “work” was hunter-gathering. Searching for food and creating shelters were the equivalents of a day-job. With no laws, kingdoms, or strong institutions, people could work however much they want. Life may have been better in that sense. But the work was hard and dangerous.
- Agriculture displaced hunter-gathering societies. Agriculture allows you to grow much more food for fewer man-hours. Agriculture existed alongside hunter-gatherer civilizations (for instance, in North America, the Aztec civilization fluorished in modern-day Mexico, while the Native American tribes continued in the modern-day US). The agricultural revolution arguably resulted in a dis-improvement the quality of people’s lives. Work became monotonous, boring, and it enabled a more stratified society. At the same time, agricultural societies allowed people to specialize in labor. With smaller numbers of people tending to food production, others could focus on science, technology, governance etc.
- The industrial revolution was the next major change in production. Machines and assembly lines got so good, that hand-crafted goods were made uneconomical. Whereas agriculture allowed for mass production of food, the industrial revolution allowed for mass production of almost everything else: cars, clothes, human capital (the modern-day education system is a product of Victorian English education system).
- Societies that underwent the industrial revolution certainly had the edge. They could create guns and weapons, expand empires. So in a sense, you could avoid industrialization, but only for a time. There were some benefits to everyday people, like cheaper clothes or whatever.
- The industrial revolution certainly had negative consequences:
- Carbon emissions began in earnest
- Labor exploitation of children and adults
- Increasing concentration of wealth to owners of capital. Hence, Karl Marx wrote his seminal work Capital.
- The cruel, abhorrent industrial-scale abuse of animals.
- Many people across the world now still work in factories or some form of blue-collar work. So these issues are very still present.
Throughout these revolutions, we see a pattern of certain forms of labor displacing others. Work in agricultural societies replace work in hunter-gatherer times. Work after the industrial revolution replaced work before that. But throughout all of it, work and jobs still exist.
Now, the present. We are living through a period of rapid technological change.
Let’s say for argument’s sake that the present era started with the invention of the computer. Since then, there’s been a variety of major transformations in global society: - computers & robots making processes more efficient. Right now, computers don’t do much for you, but they make many things you want to do much faster. Hence, almost every job in service-based economies is mediated by computers. - extremely low cost of communication & information. The internet allowed people to learn almost anything online, and communicate with almost anyone. There’s some good consequences (Wikipedia, Khan Academy, educational YouTube) and some bad (social media, polarization, misinformation). - the wide-scale availability of mobile computers (smart phones). So computers are not just used in our work lives, but our personal lives too.
And lastly, the major development of the post-computer era is artificial intelligence. By that I mean a system capable of doing “thinking” in some sense of the word. Thinking with logic and reason is what defined the Enlightenment Era, and what separates us from non-human animals. Artificial intelligence, however, could change the status quo. It could mean that very soon, we won’t be the best thinkers on this planet. In short, artificial intelligence threatens to supplant human cognitive labor. This is what underpins my predictions for the short, medium and long-term future.
In the short-term…
- Generative artificial intelligence (GenAI) will gradually get integrated into tools that we already use. Google docs, google sheets, software development, medical diagnosis.
- In some industries, like software or business operations, this will happen very quickly.
- In industries with greater inertia – licenses, government regulation – it will happen slower. For example, governments, non-profits, law, medicine are all fields where there is less incentive to change.
- In time, even the best doctors, lawyers and accounts will start to use GenAI. It saves so much time and provides such a competitive advantage, you simply must. Those who don’t will be dinosaurs within months or years. But it won’t happen overnight. ChatGPT 3 came out over two years ago. But it has barely been integrated into most people’s day-to-day lives – even computer software! It feels as if this integration should happen overnight. But in reality, it doesn’t, even for transformational technologies.
- I think this stage started in late-2022 with the release of ChatGPT.
In the medium-term:
- GenAI will do any and all job functions that require pure thinking. I’m talking accounting, law, programming, theoretical research.
- What is left for humans in this future? Plenty. Cognitive labor is only one part of most jobs in developed economies. Humans would then be better suited for:
- emotional labor. Politics, activism, sales – anything that involves persuasion. Also, care-taking, nurturing, coaching, therapy. If you take a therapist, for instance, while the AI may provide great resources intellectually, humans have some advantage in being physically present for others. We also can hype ideas up and motivate others.
- physical labor. Fixing bulbs, doing factory work, taking care of children and the elderly all involves physical labor. Teaching young kids, like I did, is partially a cognitive task, but mostly a job that requires emotional skills and physical presence.
- capital allocation and risk. CEOs, professors etc. may not do direct work themselves, but they allocate capital and therefore are responsible when things fail. More and more jobs will look like this. You may outsource work to an AI, but you’re still responsible for the outcomes.
- What is work like in this world?
- For me, it’s exciting and liberating to be freed from mind-numbing cognitive labor. Those with people-skills and the abilities to effectively use AI will benefit. Those that can’t transition may have trouble.
- The intelligence offered by AI is only useful if you work a white-collar job. What if you work in the hinterlands of America – mining or factory towns? What if you work in a shoe factory in Vietnam? Sadly, AI is not going to liberate anyone from those jobs. At least, if we confine ourselves to cognitive general intelligence.
- Some predict we will reach this level of artificial super intelligence (intelligence that exceeds the best of humans) by 2027. But I still think it will take years or decades for GenAI to truly seep into every industry in every country across the globe (just like it took decades for electrification to be truly ubiquitous).
In the long-term:
- What can the AI given enough time?
- Embodied AI (ie. humanoid robots) can do any job (physical, cognitive, emotional) that a human can. Probably better, cheaper and faster.
- AI will design and conduct full-on scientific experiments, advancing the state of the art in R&D and manufacturing.
- AI will run governments – ideally optimizing for human and animal welfare.
- What’s left for humans?
- In this theoretical world, humans would be left to do jobs that have essentially no function in society. Everyone would just be like a retiree in Florida – killing time on the golf course, in painting class, and at dinner parties.
- Realistically, there would be work to do, chiefly in politics. Who’s to say which AI system we should put our faith in? What values should the AI be optimizing for?
- We’ve had steadily increasing average incomes for decades, but poverty has not gone away, and many curable diseases take lives every year. These are not technological problems – they are political. Work would be there to distribute the benefits of AI across the world, and orient the AI so it benefits the climate and animals.
- What about work?
- Right now, there is enough wealth in the world for everyone to have ample food, shelter, and medical care. Yet that is not how the world operates. It has been the case for hundreds of years that you need to “work” to live (in the US, at least). We should expect the status quo to continue, no matter how dramatic the changes are in technology. For a work-free future, major political activism has to happen to create a 4-day work week etc. Or you can move to other countries where such futures are closer to reality.
- That said, work would hopefully be more enjoyable. It may just involve making art, playing live music, or organizing humanoid robots to do your bidding. There may be “work” to be done in colonizing the stars, melding our minds with machines, etc. Who knows?
So those are my predictions for the long-run. But let’s not forget the relevant Keynes quote, “in the long-run, we’re all dead.” In this case, we may be dead because of run-away uncontrolled AI. Or alternatively, AI may bring about next Guilded Age with astronomical wealth inequality, political assassinationsand tycoons running politics. Who knows?
Frankly, the upside of AI is just so great. And anyways, Pandora’s box has already been opened. People have seen what AI is capable of, so there’s no going back. It’s only a question of when. In the meantime, our job is to work on the issues of democratic governance, just allocation of wealth, and rights for all living creatures. The rest should sort out itself.
More blogs...
Here are some other recent posts: