"We should only make changes to Bitcoin that are absolutely necessary."

This is pushing the goalpost to an absurdly insurmountable threshold. Almost nothing is "necessary."

Take SegWit / time locks / taproot for example. It's hard to argue that they have significantly increased adoption in recent years - we'd likely be at a similar place without them. Lightning felt necessary at the time due to the scaling debate, but we haven't achieved mass adoption due to it.

However, it's very good that we have those functions available to build on top of! For example, now we have developers working on using Lightning to "glue" together different layers to effectively make them interoperable. Nobody was envisioning doing that 7 years ago.

The narrative that we shouldn't make "unnecessary" changes without "clear market demand" seems to be one made by folks who aren't willing to admit that there is always fog ahead when it comes to exploring the design space of what can be built.

We should not fear venturing forth into the unknown; that attitude would have prevented us from ever journeying this far.