I'm of the opinion that functional languages (and even just languages with strong type systems) are only useful insofar as the people using them are "cooperative" with those features. If you write ocaml as if it were Python (and if your total FP experience is a 2-week ocaml bootcamp, what else can you do?) rather than actually designing your project in a way that eg takes advantage of the strong type system to prevent invalid program states, the advantages dwindle.

But you can't write OCaml as if it were Python because the language won't let you or make it hard for you. Add to that JS's internal processes that won't let you. So, by design of the language and JS's internal structures you will write code where some classes of bugs will be absent.

I don't disagree, but I still feel that JS has a relatively low attrition rate and people there are more than competent to know the limits of their tools — that it ends up being a net positive in the long run. I wish there was a way to quantify if/when the benefits outweigh the costs.

Sure, and if you write Python as if it were Java and your total Python experience is a 2-week bootcamp, the advantages dwindle.

Have you read "Dreaming in Code"? This is a key theme in one section and I recall it was quoting Philip Eby https://dirtsimple.org/2004/12/python-is-not-java.html

It’s hurting them if the benefits are smaller than the cost of having so many employees start from 0 with the programming language. It can take months or even years to get really good with a language, especially a whole new paradigm like functional programming.