Why generics are being monopolized
Both Diamond and Kliff allude to a basic economic theory. Theory predicts that when we're on the increasing-returns to scale portion of the production function, the market tends towards monopoly (you may have heard of increasing-returns referred to as "economies of scale"). That's because at increasing returns, which ever firm is producing more faces a lower marginal cost of production, meaning they can undercut price and drive a competitor out.
Generally speaking, production functions all have the same basic shape. There are increasing returns to scale at low levels of output--where increasing production would lower marginal cost--and decreasing returns to scale at high levels of output--at some point, producing more will actually drive up marginal costs rather than lower them. If you have a bunch of factories operating at the increasing-returns portion of the production function, then you could cut costs by closing some of the factories and increasing production at the remaining factories. If those factories are owned by different firms, then you'd expect an increasing-returns industry start consolidating into fewer firms. Conversely, at the high end of output, if each of these factories is operating at the decreasing returns portion of production, then presumably the firm could cut costs by building more identical factories and reducing output at each of the existing factories. In other words, we'd expect fragmentation and new entrants into decreasing returns industries. For the most part, equilibrium should adjust until we're right at the linear-returns sweet spot in every factory.
But there is one exception. What if there's not enough demand? It is possible that the number of goods you'd have to produce to reach the linear-returns bliss-point exceeds the quantity that consumers will buy at that price. Such an industry will consolidate until there's only one firm or factory producing the good, and even then will still be on the increasing-returns portion of the production function. And once it has a monopoly, that firm can charge the profit-maximizing, monopoly price, because the threat of droping the price all the way down to marginal cost will keep out any potential entrants (since the monopolist would be producing more than the entrant and thus have lower marginal costs, the entrant cannot match this threat price).
It turns out that highly-effective generic drugs for rare conditions are uniquely profitable in this way. Daraprim, for example, only has about 10,000 users. Moreover, the number of users doesn't change much at all with price, meaning not only is the profit-maximizing price sky-high, way above costs, but that demand will never exceed the increasing-returns portion of production so there's zero potential for an entrant to compete on price even though the current price is way above marginal cost.
I got the impression in undergrad that most people think of economies of scale as a good thing. We like the feeling of saving money by buying more! But in fact it has some very very bad economic properties, as Martin Shkreli has illustrated perfectly.
What is the solution? There are a couple proposals floating around. The government could force these producers to sell at cost--there's a legal argument to be made that the federal government has this power now under the Bayh-Dole Act (Bayh and Dole do not agree). These are generic drugs, so there's nothing stopping the government (except maybe congress) from simply producing the drugs themselves, or perhaps contracting with a firm to produce them at cost. Basically, all of the proposals I've heard involve strong government action. Anyone have other ideas?
Side note: the theory is basically the same whether it's about marginal cost or average cost. Even if we are on the rising portion of the marginal cost curve, if average cost is falling (eg high fixed costs), market tends towards concentration in the same way. The largest firm can undercut the others until the competitors are unable to cover fixed costs and are forced to shutdown. Only major difference is that average cost operates on the extensive rather than intensive margin (see generally: firm shutdown conditions).
If a new entrant has to build market share gradually, then the incumbent doesn't even have to significantly drop its price. Instead it can compete by offering "volume discounts". If calculated just right, the discount undercuts the entrant while maintaining a high average price.
It is ideal for a non profit even though they would feel pressure to monopolize to further spending in other areas. They need not fear being undercut though they would need to maintain capacity to handle variable demand and readiness to switch. They could sell a smaller amount at a higher price just to sustain more than one supplier.