But here’s a catch with A.I. It’s easy to tell when a reference, or a comparison, or a sentence, doesn’t belong to a writer. Erudition and style aren’t forgeable for long; it still must be earned. As for A.I.’s sleek, space-efficient text, we’ve already grown accustomed to what that sounds like — the flat, consistent tone, the pert little summary bits, the repetitions, the impersonal and fluorescent-lit mood. Reading it, you feel you’ve been through the desert on a horse with no name.

It will get much better. Like a Nakamichi Model 500, perhaps, A.I. models will probably someday be programmed to calculate range and trajectory and to spit out rich critical prose. But as John Berryman put it in one of his “Dream Songs,” speaking of dead-on-their-feet essayists everywhere, “When the mind dies it exudes rich critical prose.” A.I. machinations can reflect the consensus, but it’s part of a real critic’s job to not go flopping along with the times, to wage guerrilla warfare on that consensus. Je suis Claude? Nix to that.

Book reviews may survive if only because, as Elizabeth Hardwick observed, publishers need praise for their new releases “as an Easter basket needs shredded green paper under the eggs.” But the breakup of the monoculture, the rise of algorithms and the flattening of taste mean that critics will never, for better and worse, have the consecrating power they once did.

Pauline Kael, Albert Murray, Lester Bangs, Edmund Wilson and Kenneth Tynan — five of my critical heroes — knew what to notice, in ways that can’t be taught or imitated, and they knew how to make their prose and their ideas stick. I’m cheered by the young critics out there, swimming in this sea without drowning in it, trying not to be cast into gaol by their creditors, and working to make certain that the last snatch of book criticism isn’t three fire emojis, two jazz-hands, a crying face and a facepalm.