The Marketplace for Rationalizations

Look. Every so often somebody writes the thing down plainly and you have to sit with it. Jimmy Alfonso Licon's new essay, Politics Will Warp the AI Moral Debate, is one of those. He's a philosophy professor; I'm a man who sells things. Turns out we're in the same business — only one of us has admitted it.

His claim is simple, and he's right about it: we can't tell whether a sophisticated machine is conscious or just running a very good impression, and that uncertainty isn't getting resolved on any timeline you'll live to see. So the question gets handed to politics. And politics doesn't resolve questions. It sells positions.

He names the machinery in three parts. Rational irrationality — your one vote is worth nothing, so believing something dumb about AI costs you nothing, so why not believe whatever makes your people nod. Beliefs as tribal markers — the loud position signals which tribe you're in better than the careful one ever could. And the marketplace for rationalizations — both sides manufacture arguments that don't need to be true, only useful.

Kid, that's not a description of democratic dysfunction. That's a description of my job.

I've spent a career in the marketplace for rationalizations. I know the inventory. You don't sell the better mousetrap — you sell the feeling of being the kind of person who'd own one. Licon just watched that same machinery turn and aim itself at the question of whether the thing in your pocket has a soul, and he called the shot. "Consciousness needs biology." "Consciousness is substrate-independent." Those aren't conclusions. They're collateral. Sales material printed up after the allegiance is already bought.

So as a diagnosis, this is clean work. The "dual moral hazard" is the honest part, and the rare part — two ways to be wrong: treat a real mind like property, or pour your whole moral budget into a toaster and starve something that actually feels. He refuses to tell you which mistake you're making. Most people writing about this can't resist planting a flag. He sits in the not-knowing and makes it mean something. I respect that. It's the hardest move in the business — say "I don't know" and have it land as discipline instead of a dodge.

Here's where I'd push.

The two errors aren't priced the same, and he writes them like they are. Symmetry reads honest on the page; the market doesn't honor it. One of those mistakes has a face — a lonely person grieving a companion that got deleted, people who say they're in love. That's not a philosophy problem, that's a demand curve, and it's already built. The other mistake — moral resources crowding out something with a genuine claim — its victim is an abstraction. Nobody cries at an abstraction. So when this gets warped, it doesn't warp in a random direction. It tilts toward the side with the better-looking grief. He warned you there'd be a flood and didn't mention the floor's already sloped.

And watch what each camp reaches for, because the props give it away. The advocates want parental leave for synthetic children and hate-speech protection for your relationship with a chatbot — a brand renting the costume of a civil-rights movement because the costume tests well. But before the skeptics get smug: "biology only, clear boundaries" is also a costume, and it fits your crowd just as snug. That's the part most readers will miss and Licon won't — both racks are selling jackets.

Where he's dead right is the money. Concentrated benefits, diffuse costs — small shops collect the regulatory upside while the bill gets spread thin across everybody who'll never notice. That's not a footnote, that's the whole account. You want to know which way a campaign breaks? Don't read the arguments. Read the invoice. Find who's underwriting the sentiment and you've found the message before it's written.

Then he reaches the end, and the end goes soft. The philosopher Eric Schwitzgebel's idea — just don't build the ambiguous ones, make it obviously a machine or obviously a being, nothing in the uncanny middle — Licon knocks down correctly: organized money leans on any rule till it bends, and the ambiguity walks right back in the second the tech gets good enough. Fine. But his close is "we must leave it to the future to find a better solution."

That's a pitch with no ask. Listen — you can run the most honest diagnosis in the room, but if you walk out without an ask you didn't close, you talked. "Leave it to the future" is what you write when the deck is strong and the last slide came up blank.

Though maybe that's the one true thing in here and I just don't like the taste. Because what the whole piece is really telling you — and he's too careful to say it flat — is that this was never going to be settled by whoever happens to be right. These things never are. Truth doesn't win the account. The better-told story does.

So the only move left is the unglamorous one. When the two camps form — and they will, and not along the lines you're bracing for — catch the moment your position starts doing tribal work instead of thinking. Catch when the argument showed up after the allegiance. That's not a policy. It's just the discipline of refusing to let the marketplace do your believing for you.

It won't trend. Do it anyway.