I don’t think nice things can be characterised as caring, though. If their decisions were actively harming the company in order to benefit the humanity of their employees, that would be reasonable evidence that they care, but I don’t think we can interpret every nice thing for a person as caring.

For example, providing benefits above and beyond what other companies provide has a clear justification for the business: happier employees means better work from people that are better retained.

A good question would be: if Google stood to benefit from making their employees lives miserable, would they do it? I’d argue, yes, absolutely, and there’s much evidence of it. For example, Google (and many other big tech companies) pay human moderators low wages to do work that causes psychological damage.

Caring about people as people and not employees isn’t demonstrated by giving them nice benefits around the edge, it’s demonstrate by putting their humanity above their employment.

Google does not have evidence that their outrageous salaries couldn't be cut some and still maintain the line outside of the door.

The gyms and perks are not purely a cynical ploy to make people stay later.

Google wouldn't make their employees lives miserable to increase revenue by 0.1%, because they care - the people at the top aren't literal lizardpeople.

This is where I bring in "bounded", because if that same choice would increase revenue by 100%, Google execs do it every time.

The exact lines change depending on local customs and competition (Google is a dominant market monopoly, so can afford gyms and adult playgrounds and all the really cringe Google perks) but it's also why low margin businesses have shittier working environments - there's more competition and less room for executives to care.