And... that's right, rol. There is only so much they can use up. Andrea taps that, too, just now. It's a major part of why increasing production efficiency hasn't resulted in an increasing workforce.
To supplement some of what 's said above, some of y'all seem to be forgetting demand and market saturation. You can't just infinitely ramp up your production and expect things to turn out well -- if no one's buying, or no one's willing to buy for more than it costs you to make, pumping out more stuff does sod all except lose you money and the ability to support what workers you did. If it's reached the point where the market just doesn't need the gubbin(s) your factory is tooled to build, or doesn't need nearly as much of it, or has stopping growing as quickly and no longer can support you increasing production as quickly, again, pumping out more does sod all.
Sometimes retooling is an option but there's a whole host of new variables that brings in, and it all it does is kick the can down the road a bit. The human population and especially those with the resources to buy stuff by and large hasn't been growing faster than our improvement in making crap -- that's one (of the myriad) reason(s) manufacturing et al caps, just hard, flat caps at a point and stops being able to keep up with increases in potential new workers. It certainly wouldn't keep up if we had been increasing workforce and subsequent output hand in hand with production.
It's, like. Materially, workforce, etc., etc. We could ramp up vehicle production, just as an example. We could ramp up vehicle production so much we would have literal mountains of vehicles (well, more of them, anyway) with no one to sell to and nothing to do with them, sitting there rotting, and then what? Broken window theory and the resulting mess of rusting steel only does so much, and there's other issues to consider as well (limited resources, mostly).
Or to point to something RP said, they were wrong when they flipped it. Flat Production Amount that we need a certain number of people to support is exactly what's going on, because more production past a certain point is less than useless.
... all that said, productivity gains isn't what's been causing QoL life loss in the areas manufacturing et al changes are hitting the hardest. Not adapting to it is what's causing that. Not retraining, not moving, not using the existent infrastructure (what there is) and expertise to dig out new markets, not having prepared for what was coming before the work started leaving. Far from all of that is even remotely the workers' fault (again, note everything I've mentioned previously re: funding, and add on to that the next to zero incentive businesses that more or less own towns have to future proof them), but... the QoL drop could have been avoided, or at least significantly mitigated. "All" it would have taken was everyone that had been working to reduce the funding and resources aimed at fixing the problem dropping dead, or at least out of politics, about fifty or sixty years ago.