I knew a landlord in East Hollywood who always brought pears.
Good pears too. The kind that come wrapped individually in tissue paper, like they’re some kind of ceremony. He’d knock on the door, and before you could get your pants on, he’d be standing in the hall with a little basket, smiling the smile of a man who had something to say and wanted the preamble to soften it.
“For you,” he’d say. “I got too many.”
Nobody has too many pears. Nobody accidentally ends up with four Bartletts and a D’Anjou and decides to distribute them to their tenants. But you’d take them, because you weren’t a monster, and because his smile was doing something that looked like warmth, and because you knew the script. He’d linger. He’d talk about the weather, his wife’s knee, how gas prices were killing him. And then, just before leaving — always just before — he’d mention the rent.
Going up forty. Going up sixty. Going up whatever he needed that month.
You’d be left standing there with a pear in your hand and no good way to throw it, because that would be rude, and because he was already halfway down the stairs.
I thought about him this week.
OpenAI published a thirteen-page policy paper. I’ll say that again, because the number matters. Thirteen pages. That’s the whole thing. Thirteen pages, double-spaced probably, explaining how the machines they sell are going to eat the American workforce, and what they propose to do about it.
Their proposals are beautiful. I mean that. If you slid them across a bar on a cocktail napkin, I might have bought you a drink. A public wealth fund. Higher taxes on companies that replace workers with AI. A four-day workweek, funded by something they called “efficiency dividends” — as if the future were going to be so abundant that the crumbs alone could feed us. Government programs to shepherd displaced workers into “human-centered” jobs. The whole fruit basket.
And on the same day — this is the part where you start laughing or start drinking, and I did a little of both — The New Yorker dropped seventeen thousand words on Sam Altman’s history of lying.
Seventeen thousand words. Thirteen pages. That’s the ratio of truth to pears in this business.
Ronan Farrow and Andrew Marantz wrote it. Farrow is the guy who put Weinstein in a cell, so when he spends seventeen thousand words on a man, he has receipts. And the receipts tell a simple story: Altman tells people what they want to hear, then walks into the next room and does the opposite.
He told lawmakers he wanted federal regulation of AI. Then he privately lobbied to kill the specific bills that would have provided it. A state legislative aide in California described OpenAI’s behavior as “increasingly cunning, deceptive.” Last year the company subpoenaed the supporters of a California AI safety bill — not because it was suing them, but because, as one of the targets put it, they wanted to scare them into shutting up. Subpoenas as a muzzle. Pre-litigation pears.
That kind of move requires a specific confidence. The confidence of a man who has figured out that saying the right thing in public is free, and doing the opposite in private is where the money lives.
So now he wants to talk about a four-day workweek.
Here’s what the thirteen pages don’t mention: the workers being “saved” by this benevolent redistribution are the same workers being deleted by the product. The math is elegant. I replace you, I tax myself a little for the privilege, and then I pay you some of the money back as a stipend to find “human-centered” work — a phrase so vague it could mean flipping burgers or writing haikus about your own obsolescence. Probably both. Probably the burger place wants the haikus for their window.
I worked at the post office for years. I know what corporate kindness looks like when it’s printed on company paper. The wellness memos. The retirement seminars. The free donuts on the day they announced the route cuts. You learn, if you stay long enough, that a long memo is usually an apology for a coming injury. The longer the memo, the deeper the knife.
Thirteen pages is a long memo.
And here’s the part that grinds on me, sitting with this glass of something brown and the afternoon sliding sideways: there are probably people on that policy team who meant it. Who stayed up late with coffee and good intentions, writing paragraphs they believed in. A guy from MIRI said as much in the article — he figured some of them actually cared. Then he said, almost as an aside, that those people were probably about to find themselves in “the position that many previous people at OpenAI have found themselves in” — which is a polite way of saying they were about to get the pears, and then the eviction.
That’s the part nobody writes about. Not the villains. The believers. The technical safety researchers who thought they were drafting a blueprint for the future, when they were actually drafting an alibi for the next lawsuit. The kids who wanted to save the world and ended up working for the man who subpoenas people who want to save the world.
I think about that landlord sometimes. He retired eventually, I don’t know where. I hope his pears went bad on him. I hope he opened a cupboard one morning expecting tissue paper and smelled something rotten, and understood, for once, what he smelled like to everyone else.
But that’s not how these things usually go.
Usually the pears stay good all the way down. Usually the smile holds. Usually you get the paper with the pretty ideas about the four-day workweek, and you nod, and you read about efficiency dividends, and you don’t notice the part where the rent just went up sixty dollars and the locks get changed on Thursday.
And somewhere in a nicer part of town, a man is writing another memo about how much he cares.
Source: OpenAI made economic proposals – here’s what DC thinks of them