I just don't believe there's some cliff of automation effectiveness where it goes from a good thing to a bad thing. Greater productivity always means making more for less.

Only if we think if only in first-order effects. But there are always second and third order effects.

Environmental costs is one of these.

The desirability (from a societal standpoint) of having more of X (whatever X is, e.g. cheaper more plenty fast food for a controversial example) is also another matter.

The impact of the jobs that replaced the deprecated ones is another issue (a third order effect).

And lots of other considerations...

As the price of labor trends downwards, some cohorts are going to have a harder time selling their labor to pay for necessities: food, shelter, healthcare, etc.

A good amount of these problems could be mitigated by a strong welfare system and/or basic income.

The price of labor tends to be lowest in the lowest-productivity areas. And there are good theoretical reasons for that to be the case -- high labor prices tend to cause productivity increases as a way to avoid paying them, and productivity increases increase the value of labor, which would tend to lead to higher labor prices.

When everybody was a pitiful subsistence farmer, barely producing more than they needed to eat themselves, the price of labor was very low. And where that's still true... the price of labor is still very low.

Can you point to a current or historical example of the price of labor going down as productivity rises?

We might be using terms in a different way here. This statement is true in one sense, and untrue in another sense.

When there is higher productivity, an hour of labor accomplishes more work. Employers are getting more bang for their buck - an hour of laborer pay accomplishes more work for the employer. (i.e. the employer is getting more value)

On the other hand, because it's possible to accomplish more work with less labor, the demand for that type of labor drops. This leads to depressed wages (i.e. the employee is getting less value). A good historical example of this was at the beginning of the industrial revolution, when power-looms caused weaver-labor to become nearly worthless https://en.wikipedia.org/wiki/Power_loom#Social_and_economic...

Many of the displaced workers became Luddites who wanted to smash the machines which put them out of work. You would think that being relieved from working would be a blessing, but not in their circumstance. Because the weavers' labor was so sharply devalued, they found it extremely difficult to sell their labor for enough money to eat.

The short-term effect isn't the whole story, of course. Textiles were incredibly cheap which did increase demand, and eventually led to the creation of (low-paying) jobs in textile factories. But even where things turn out fine in the long-term, the short-term disruption is not to be underestimated. Widespread unemployment due to power-looms caused riots.

> When everybody was a pitiful subsistence farmer, barely producing more than they needed to eat themselves, the price of labor was very low.

To the contrary, the price of labor then was incredibly high. As you say, serfs spent the majority of their time simply generating their own subsistence. Feudal lords were entitled to a portion of the serfs' labor, but it only amounted to a day or two out of the week. Consider the legions of subsistence farmers it would take to support specialized full-time jobs.

Once industrial agriculture started taking off, food was suddenly cheap and plentiful. This made sustenance much less expensive and freed up labor for other non-farming purposes.

> high labor prices tend to cause productivity increases as a way to avoid paying them

This I agree with! In fact I believe it expresses the same point that I'm trying to convey: increasing productivity means employers pay less for labor.

Witness the rise of self-checkout stations at grocery stores, as well as kiosks and automatic-burger flippers at fast food restaurants. The whole point of these innovations is to decrease the demand for such labor.

And a good amount of the advantages would be mitigated too.

Automation is good for everybody if we rethink our assumption that wages are the best way to get everyone money to live on.

Many capitalists however are religiously attached to that idea.

The problem is what to do about people who live in locales where there aren't a lot of sales? So they don't get much money and don't increase sales. A catch-22.

That's one issue that crops up in office automation, but not necessarily the biggest one. General tools like word processors and speaadsheets remain just that: tools. But once office functions gain special purpose software, it often locks in strong assumptions about workflows that translate into more, rather than less, "sorry but I can't do that."

Which just solidifies the demand for people to build more such software for other use cases.

> “I’m just really surprised that anyone would say that this is the most automated industry,” says Rader. If someone orders a plaque from his company, a machine engraves it. But a human lays out the appropriate design using a computer, selects the materials, lays the metal in the machine, assembles the plaque, and packages the plaque when it’s done.

The barrier to entry is much lower and the skill-level required to do a moderately competent job has plummeted.

To me that would be the final nail in the coffin. But then again, maybe the powers would be shifted from the person who runs the machines to the person who orders the engravings or the person who consults with promotion companies or to the one who opens up the possibilities of more complex engravings.