There are numerous configurations, networks, and commercial structures that can be created for NFT offerings that can mitigate "problems" described in the OP and accommodate for different tradeoffs.
A simple retort would point to the many alternative NFT networks that don't use POW-based consensus algorithms or the NFT offerings that enable value add offerings very differently than purchasing art or baseball cards.
Evaluating the capabilities of NFTs against the value offerings of something like rare art is an apples to oranges comparison. Of course the market for digital assets offered via NFT is immature and speculative; this is a new market, based on an alternative technology paradigm, that has a long way to go before it settles into a more usable and valuable structure.
Ultimately, the idea of using key-pairs pegged to widely accessible peer-to-peer public networks as a mechanism for tracking ownership of digital (or near digital or at times even physical) assets is incredibly novel. It turns the conventional model of third-party hosted digital assets on its head and enables really interesting mechanisms of distribution, ownership, access, and value consumption and creation that does not compare well with traditional mechanisms. And in saying it doesn't compare well I mean to acknowledge its limitations and its potential at the same time. However, the critique in the OP is pretty bland and doesn't seem to acknowledge the full scope of the situation.