My argument for owning Eth is that it's like buying real estate (yes, I'm making another loose analogy). Why does a person invest in land? Because they expect something to flourish in the vicinity of their land, and they'll be able to capture some of that increased productivity in the form of rent (or appreciation).
Ethereum is a programmable ledger, and Eth is its unit of capital. A programmable ledger is a tool that can be used to create things like a money supply the author describes. The next question here is "why would someone prefer to use an analogous system created on Ethereum over whatever exists today" and that answer can be some combination of transparency, usability, and legitimacy. If we want something like the elastic-supply currency the author describes, we can create one for ourselves. We can do so without providing limitless power to a centralized authority (i.e. the Fed). To some, the resulting currency will be more legitimate as a result. (checkout Rai or Liquity for examples of elastic/"algorithmic" stablecoins on Ethereum).
In this view, today's value of Eth is largely speculative. On the other hand, this gives a way to fund development. New protocols like the ones I mentioned are able to capture the value they create only because there's already so much $ in this ecosystem, and that money's eager to pay a 0.1% fee for the benefit of this new product.
There's a lot of circular activity going on, sure. Deflationary rewards, leverage, etc. If you haven't seen an 80% drop in crypto yet, be ready for it (seriously). But as long as the influx in capital facilitates some level of actual new products -- as it does appear to me to be doing -- then what really is the issue here? Maybe we can be satisfied in viewing it as a rather complicated machine for capital allocation? My biggest worry is that impending 80% drop: crypto is notorious for its volatility, but still each bull run draws in newcomers who think we're past all the drops and find themselves surprised when they lose 80%.