Working on my own kind of metaverse I've grappled with this issue, and that is why I hope to use it to improve physical space endeavors as opposed to replacing them. This really started for me in two places: the newer one was when I pitched to the department of education for a grant program a software I called Meta-Education Environment for Simulation and Gaming (MEESG), aimed particularly at assisting in training of high-risk vocations (such as high voltage-systems repair)... but after I thought about it, I realized that the earlier inspiration was the VBS simulation I trained in the military and how I noticed the benefits directly (later leading to many hours of Arma2/Arma3 playtime, the civilian version of VBS).

So the main fight I see in the future is those who think (similar to the article) all this immersion effort is vapid and superficial. I don't think it is, and for more than just training, but also for rapid iteration in a simulated physical space that doesn't waste actual physical resources until a better product is developed. Ergo, I feel there is hidden value in the virtual space yet untapped in the wider market for combining fun and relaxation with teaching valuable things about the real world.

Just for example, I have been adding my local flora/fauna and edibility properties and medicial properties to one of my gameworlds recently, which could help me accelerate my learning of that particular real world thing but also make the gameworld more fun and interactive.

An even darker take, if some renegade member of an advanced alien race does decide to spend a lot of resources on exploring the universe, and gains the means to do so, the other members of their species may very well plug them into a simulation while they are sleeping and make them think they are exploring the cosmos.

Unless a species is completely post scarcity, there is a strong ethical argument to be made that this is the correct thing to do, versus spending a lot of resources on "mere exploration",