Yes, if you had one caretaker looking after thousands of children, quality would be poor. But that doesn't mean the optimal number is 1. A professional caretaker looking after a manageable number of children can certainly outperform an amateur looking after one or two, and a facility with multiple specialized caretakers can outperform the single professional caretaker.

You don't want to minimize students per teacher, you want a healthy number of students per teacher. Class sizes are not optimal at 1. Below some minimum class size (which varies by age group) there is no benefit to further reduction, and sufficiently low numbers can be harmful. That's to say nothing of the additional cost of that labor to achieve such faculty ratios.

You've gone from efficiency and economies of scale, to a "professional" outperforming an "amateur." Raising a child is not like making a widget. Endless studies [1] demonstrate that more early non-parental care leads to worse outcomes in just about every single way - worse behavior, health, attention span, long term higher likelihood of police encounters, and much more. An interesting one is that children who spend extensive time in daycare even end up less socially competent which is quite interesting since it runs contrary to one of the typical arguments in favor of daycare. But it's also not surprising if you think about it, because at home a child is getting vastly more attention and interaction than he would in daycare.

And this is especially significant because that's just speaking aggregately. Obviously not all parents are created equal, but it turns out that even bad parents tend to be better than non-parental care, especially early on. If you isolated it only to active, highly involved, parents - the results would be exponentially better than they already are.

[1] - https://search.brave.com/search?q=long+term+outcomes+of+dayc...

Amateurs regularly outperform professionals in schooling (they seem to perform somewhere between "at least as good" to "decently better" on average), and studies in the 80s found that 1:1 tutoring with mastery learning is wildly more effective than normal classes (with the average tutored student performing at the 98th percentile of control students).