Michael Sandel has another reactionary book coming out, which based on this discussion, appears to be classic Sandel. (As one of my friends said, Sandel's basic argumentative style is, "Hey, you know all your untutored, unreflective prejudices? You should roll with those. I'm from Harvard.")
Based on this discussion, you might wonder whether Sandel is more worried about whether meritocracy is bad in itself, or instead whether the problem is that we act like we're meritocratic but in fact we mess it up.
One of the big problems any organization, including entire webs of social cooperation, face is that when people try to distribute some good (status, property, income, admission to Harvard, Harvard professorships) on the basis of some feature or action (hard work, intelligence, moral virtue, civic virtue, wokeness, publishing prowess, insight, productivity, innovation, or whatnot), then people will try to fake the signal to get the good. When we start measuring something, people fake the measurement.
For instance, it seemed like a good idea for elite schools to select altruistic, public-spirited students. But once students realize schools were using volunteer hours or "I started a non-profit when I was 15" as an admissions criterion, then everyone starts volunteering in order to fake the altruism signal. The signal becomes diluted and so we get an arms race to produce an ever stronger signal.
These sorts of problems might show that we suck at distributing goods according to merit. They do not, however, show that actually distributing things according to merit is bad.
Note that I'm not a meritocrat. My point is that as philosophers, we must distinguish between whether a principles of right action (which identify what is right or wrong) from the institutions we use to realize those principles. For instance, it's good to have Pareto-optimal outcomes, but that doesn't mean we should have the Department of Pareto-Optimality aim for it. (And if we do and they suck, the problem is the department, not the principle.) Happiness is good, but you might do better not aiming directly for it. Similarly, distributing certain things according to merit might be right, but then it's an empirical question which mechanism for distribution successful tracks merit. It might be that directly meritocratic structures suck at producing meritocratic outcomes. The problem is the structures, not the principle.