The robots won’t have all the jobs. And the demand for human labor will increase.
There will always be some jobs humans can’t do. It’s not that there’s something magical about humans and the human mind. It’s just that there are certain jobs that are so complex and involve such human emotional intelligence and human interaction, that any machine that could do this instead of a person would have to be a person themselves. I might trust Commander Data to be my kid’s elementary school teacher. But that’s also because I would consider Commander Data to be a person. But there would also be little reason to mass produce Commanders Data to be elementary school teachers, as that would amount to little more than slavery. A mind is a mind, regardless of the substrate. Forcing a mind to work for you is a moral abomination, regardless of whether that mind is flesh or silicon.
As automation has increased over the generations, the demand for human labor has increased. The fields whose services have increased in price high above inflation are non-coincidentally those with the highest amount of unavoidable human labor. Think medicine, higher education, and home construction. Automation generates vast wealth. People who profit from highly automated industries then have more money to spend on things. There’s more money in the economy to support the labor-intensive industries. But automation can’t meaningfully decrease the cost of producing them. So the wealth generated in low-labor intensity industries goes towards bidding up the cost of the goods and services produced in high-labor intensive industries.
Or another way to look at it. Automation is deflationary. Whenever the production of a good or service becomes highly automated, the cost that good or service tends to go down. There’s a reason the idea of a walk-in-closet would have been considered absurd to your ancestors. When people spend less money on the automated goods, they have more money to spend on the labor-intensive goods.
Or, a third perspective. A reasonable assumption is that future automation will look like past automation. Yes, automation can be disruptive on an individual level. If you’re 55 and your entire career specialty is automated away, you’re going to be really hurting at a personal level. You just don’t have time to retrain for a new career field, and medically you may be unable to. But as a whole, people move into fields that have high need for workers. We have a higher labor force participation rate than we did 200 years ago, despite only a single-digit percentage of people needing to work in agriculture now. Wave after wave of automation has failed to result in the predicted mass employment and immiseration of the populace. And every time we’re told that “this time is different,” it turns out to be no different than the previous times. The people telling you that this round of automation will be completely different from all the others are the same people profiting from the current AI bubble.
The automation that AI is promising (but not necessarily delivering) is fundamentally different than the automation that came before.
Remember; the luddites were right but their industry was small enough that the displaced labour could be absorbed by other industries.
Not only is AI affecting almost every artistic and white-collar industry, but the cap-ex barrier to entry is way lower than for any other automation effort in the past. No need to buy expensive machines, or create whole new production lines just to test it out. Computers used to take up an entire room to do the work of a handful of people. If you can increase productivity but there is no associated increase in demand, then what you get is layoffs.
The amount of workers that this has the potential to displace far outstrips the industry/economy/society’s ability to replace with “new careers” (that we’ve yet to see materialize). And I challenge your assertion that automation has resulted in increased demand for human labour, do we significantly less unemployment on average? Over the last 70 years (in the USA for ex) unemployment has been trending up.
What we have seen is a total gutting of employee bargaining power.
Why do you assume these all need to be new careers? Why do you assume that we can’t expand existing careers? It’s happened in the past, it can happen again. Agriculture went from employing the majority of the populace to 2%. We found jobs for everyone.
There are many professions that have immense latent demand that people simply cannot afford. Really any industry that involves a lot of human labor. People want more education than they can afford. People want more healthcare than they can afford. People want more childcare, private tutoring, home cleaning, personal trainers, life coaches, financial advisors, and on and on. Think of the retinue of assistants and employees the wealthy employ. Now imagine the number of people who can afford those services drastically expanding. We don’t even need to necessarily invent new careers. There’s plenty of latent demand already. Those masses of displaced agricultural workers? Most of them found jobs in fields that already existed.
Over the last 70 years (in the USA for ex) unemployment has been trending up.
This is false. I’ll ignore the employment rate and focus on labor force participation rate, as unemployment doesn’t count people who are long-term unemployed and have given up working. Labor force participation is a better metric here.
Labor force participation has gone up and down, corresponding with changes in demographics. Despite generations of technological change and automation, we’ve always found ways to employ the excess labor. Human labor is always the ultimate bottleneck. There’s probably enough latent demand for human labor to employ many multiples of our current population.
The robots won’t have all the jobs. And the demand for human labor will increase.
There will always be some jobs humans can’t do. It’s not that there’s something magical about humans and the human mind. It’s just that there are certain jobs that are so complex and involve such human emotional intelligence and human interaction, that any machine that could do this instead of a person would have to be a person themselves. I might trust Commander Data to be my kid’s elementary school teacher. But that’s also because I would consider Commander Data to be a person. But there would also be little reason to mass produce Commanders Data to be elementary school teachers, as that would amount to little more than slavery. A mind is a mind, regardless of the substrate. Forcing a mind to work for you is a moral abomination, regardless of whether that mind is flesh or silicon.
As automation has increased over the generations, the demand for human labor has increased. The fields whose services have increased in price high above inflation are non-coincidentally those with the highest amount of unavoidable human labor. Think medicine, higher education, and home construction. Automation generates vast wealth. People who profit from highly automated industries then have more money to spend on things. There’s more money in the economy to support the labor-intensive industries. But automation can’t meaningfully decrease the cost of producing them. So the wealth generated in low-labor intensity industries goes towards bidding up the cost of the goods and services produced in high-labor intensive industries.
Or another way to look at it. Automation is deflationary. Whenever the production of a good or service becomes highly automated, the cost that good or service tends to go down. There’s a reason the idea of a walk-in-closet would have been considered absurd to your ancestors. When people spend less money on the automated goods, they have more money to spend on the labor-intensive goods.
Or, a third perspective. A reasonable assumption is that future automation will look like past automation. Yes, automation can be disruptive on an individual level. If you’re 55 and your entire career specialty is automated away, you’re going to be really hurting at a personal level. You just don’t have time to retrain for a new career field, and medically you may be unable to. But as a whole, people move into fields that have high need for workers. We have a higher labor force participation rate than we did 200 years ago, despite only a single-digit percentage of people needing to work in agriculture now. Wave after wave of automation has failed to result in the predicted mass employment and immiseration of the populace. And every time we’re told that “this time is different,” it turns out to be no different than the previous times. The people telling you that this round of automation will be completely different from all the others are the same people profiting from the current AI bubble.
The automation that AI is promising (but not necessarily delivering) is fundamentally different than the automation that came before.
Remember; the luddites were right but their industry was small enough that the displaced labour could be absorbed by other industries.
Not only is AI affecting almost every artistic and white-collar industry, but the cap-ex barrier to entry is way lower than for any other automation effort in the past. No need to buy expensive machines, or create whole new production lines just to test it out. Computers used to take up an entire room to do the work of a handful of people. If you can increase productivity but there is no associated increase in demand, then what you get is layoffs.
The amount of workers that this has the potential to displace far outstrips the industry/economy/society’s ability to replace with “new careers” (that we’ve yet to see materialize). And I challenge your assertion that automation has resulted in increased demand for human labour, do we significantly less unemployment on average? Over the last 70 years (in the USA for ex) unemployment has been trending up.
What we have seen is a total gutting of employee bargaining power.
Why do you assume these all need to be new careers? Why do you assume that we can’t expand existing careers? It’s happened in the past, it can happen again. Agriculture went from employing the majority of the populace to 2%. We found jobs for everyone.
There are many professions that have immense latent demand that people simply cannot afford. Really any industry that involves a lot of human labor. People want more education than they can afford. People want more healthcare than they can afford. People want more childcare, private tutoring, home cleaning, personal trainers, life coaches, financial advisors, and on and on. Think of the retinue of assistants and employees the wealthy employ. Now imagine the number of people who can afford those services drastically expanding. We don’t even need to necessarily invent new careers. There’s plenty of latent demand already. Those masses of displaced agricultural workers? Most of them found jobs in fields that already existed.
This is false. I’ll ignore the employment rate and focus on labor force participation rate, as unemployment doesn’t count people who are long-term unemployed and have given up working. Labor force participation is a better metric here.
Labor force participation has gone up and down, corresponding with changes in demographics. Despite generations of technological change and automation, we’ve always found ways to employ the excess labor. Human labor is always the ultimate bottleneck. There’s probably enough latent demand for human labor to employ many multiples of our current population.