Yeah, but the question is, where are these economic gains from automation and AI going? They don't just go into the ether - they'll accumulate to whoever financed the automated whatsit, and rightly or wrongly that person will not willingly share their vastly increased income with a superfluous zero-leverage former workforce. This is exactly what is happening now as massive productivity gains over thirty years have failed to translate into an increase in real wealth distributed across the population. The technological solution creates a social problem.
From the "What If Technology Is Destroying Jobs Faster Than It Creates Them?" one of the links of the article:
> It’s beginning to look like we might have entered a two-track economy, in which a small minority reaps most of the benefits of technology that destroys more jobs than it creates. As my friend Simon Law says, “First we automated menial jobs, now we’re automating middle-class jobs. Unfortunately, we still demand that people have a job soon after becoming adults. This trend is going to be a big problem…”
And this is of course true, the only thing that makes software cheaper is the one resource it can actually save on : people. I would even argue that, generally, when software replaces a person, it is actually less efficient in all but one metric : it generally uses more resources than a person would use for the same work.
So I'm arguing that mostly the efficiency gains from technology are negative, but you're trading expensive resources (people) for cheap ones (power, computers, robots ...).
A few more arguments for this :
1) machines are more precise, but require better inputs, which require more resources, and what they can't deal with is generally not recycled, as that would destroy the value of those machines.
2) the human body is surprisingly efficient (especially knowing how general it is in function. Usually the more generally useful something, the less efficient. Humans are among the most efficient animals and are vastly more general). Our body beats oil based energy, and beats coal based energy by a large margin (input in watts versus output in work).
3) saving the expense of keeping the human alive is useless ... unless, of course you intend to kill/starve/ignore till they're dead/... said human. You're just creating a negative externality.
4) as the past centuries have pointed out, saving human labor is completely useless if they're your market. The last French king had a much less comfortable life than a homeless man has today, with few exceptions : servants and available space. But seriously, even the food that was served to that king, I wouldn't touch it (I shudder to think what medieval kings ate). His toilet is only slightly better than a hole in the ground. Sheets, beds, ... all are cleaner, more functional and a LOT softer today. And about cleaning ... well people actually clean themselves today. I shudder to think of the smell of Versailles in the 19th century.
The argument "but life never gets worse in history" has plenty of counterexamples, the biggest one (imho) the end of the Roman Empire, which took half a millennium to recover from. But there were plenty, say, the muslim conquests of Northern Africa would be another example. Started out with >50% of the population pretty comfortable (e.g. running water, heating, public fountains, roads, working medical infrastructure, surgical procedures available to middle class and even a significant part of the slave class), working travel infrastructure, life expectancy at ~57 years and ended with tribes who effectively lived in the stone age (average life expectancy at one point under 20), and a few lords who lived at a comfort level that was at best what it was at the beginning of the iron age. They would remain locked in this situation for more than a millennium. Some pockets of technology remained, like Cairo, which retained it's Christian institutions up to this day. But technology can't compete with free labor, so it couldn't expand despite many, many tries.
