1. The author fails to explain clearly that what we call mining "mining" is a distributed form of transaction processing, in which transaction fees are earned via proof-of-work. Currently most of the transaction fees come from block rewards, but transaction fees are in fact set by competition based on the marginal cost of computation.
2. By design, Bitcoin's block rewards have always been meant to be temporary, as explained by Satoshi Nakamoto in his/her/their paper. I mean, the rewards are halving on a fixed schedule, and everyone knows that eventually there will be no more rewards. The only purpose of the block rewards was -- and is -- to provide a temporary incentive for jump-starting Bitcoin's transaction processing network.
3. As the rewards go to zero, Bitcoin transaction fees will fluctuate. They will be determined by competition based on the marginal cost of computation. My expectation is that Bitcoin transaction fees will go up over time.
4. Many transaction processors, i.e., "miners," are convinced that Bitcoin eventually will have a market capitalization similar to that of gold (~$12T, give or take), so they are hoarding their share of block rewards. Many investors in these operations are convinced too. Some in fact are willing to lose money processing transactions to get the block rewards.