Enshittification is inevitable for all free services (services as in with a server component).

No, it is not that bleak. It is only inevitable when there is an active push for a short-term maximization of user base monetization (which is very much in the nature of VC). It can usually be avoided with products that are wholly under the ownership of all users (such as a cooperative or a government-provided service) or - only if one is lucky - with products of financially independent private enterprises under vaguely benevolent and unhurried leadership (such as Steam, to some extent)

Monetary needs and all that. If it’s a startup with VC then there is either not enough people paying or not enough private users supporting by other means like bug fixing, support, etc. Or greed by VC.

Well, VC is greedy by design. A VC-funded business will never be optimized for longevity, a good product or happy customers. They may achieve those things en passant, but they’re never the objective.

For example: Any case of “there is not enough people paying” can also be rendered as “the scale and moving speed of the business is way off”.

Why do people host LLMs at home when processing the same amount of data from the internet to train their LLM will never be even a little bit as efficient as sending a paid prompt to some high quality official model?
inb4 privacy concerns or a proof of concept this is out of discussion, I want someone to prove his LLM can be as insightful and accurate as paid one. I don’t care about anything else than quality of generated answers

If you ask other people for their reasoning and opinions, it doesn’t really make any sense to put something “out of the discussion”, does it? :P

But no, if you have no qualms about sharing your innermost feelings, sexual preference or illegal plans with those that have an explicit desire to exploit that information then there is little reason to attempt something as complicated and wasteful as self-hosting your own LLMs.