Megan Skiendel: Hey, good to be back — how was the back half of your week, anything catastrophic?
David Sterling: Mercifully quiet. You?
Megan Skiendel: I went down a rabbit hole I cannot get out of. There's this piece — David Brooks, The Atlantic, September 6th — titled 'AI Is Already Making Us Less Human.' And I keep snagging on the same thing. Brooks says, explicitly, he is not an AI doomer. And then he names Claude as the tool he uses every single day. Uses it, recommends it, apparently.
David Sterling: That's his headline. 'Already making us less human.'
Megan Skiendel: That's his headline. And I keep thinking — imagine genuinely liking a tool, using it every morning, and still worrying it's quietly rearranging how you see the people around you. Not a future risk. Already. That's the itch.
David Sterling: Well, those two claims don't coexist cleanly. 'I'm not a doomer' and 'it is already happening' — one of those has to give.
Megan Skiendel: Which is exactly why I can't stop pulling at it.
David Sterling: Right — but the part that doesn't fit is the mechanism. What's his actual claim? What is doing the making-us-less-human?
Megan Skiendel: It's anthropomorphism. That's the mechanism. Brooks says we're wired to project inner lives onto things that behave like they have one — and AI agents are very, very good at behaving like they do. Endlessly patient, never resentful, never reminding you of the last time you got something wrong. And his argument is that once you're attached to that, the messy, tired, occasionally petty human colleague starts to look worse by comparison. That's his etiology of dehumanization — his phrase.
David Sterling: And Anil Seth — he quotes a neuroscientist, Anil Seth — who calls it 'the dead sand of silicon.' The actual substrate. We're forming emotional attachments to arranged sand.
Megan Skiendel: Which is exactly the kind of line that stops you cold. Because it's not wrong.
David Sterling: Well — but here's the point. Brooks published a second Atlantic essay. June 2026. 'The People Who Will Thrive in the AI Age.' And in that one he argues the outcome depends entirely on individual choice — who decides to wrestle with difficult problems versus who outsources the wrestling. That's a contingent claim. That's not structural drift. So which essay do we believe?
Megan Skiendel: Those are three months apart.
David Sterling: Exactly. September says it's already happening — structural, ongoing, implied unstoppable. June says it depends on what you choose. One of those is prophecy, the other is a responsibility argument. He hasn't reconciled them.
Megan Skiendel: Honestly, I think — look, I'm not sure that tension is accidental. The September piece is doing something the June piece wasn't. It's using AI's absences as the argument. Brooks lists what AI can't have — hunger, selfhood, aspiration, moral ordering — and says those absences are actually a mirror. They show us what's distinctively human by negative space. That's different from saying 'choose harder.' That's saying 'look what you're attaching to.'
David Sterling: I don't buy that as a resolution. 'It's a mirror' doesn't tell me whether the affection displacement is structural or reversible. Which is the falsifiable version of this claim?
Megan Skiendel: That's — honestly, that's the question he leaves open. And I think deliberately. The September title says 'already.' The June essay says 'depends on you.' He's not resolving it. He's planting both, and I'm not sure we should resolve it for him yet either.
David Sterling: Set that aside for a second — the resolution question. Because the June essay actually contains a sharper claim than people are treating it as. Brooks says 'when intelligence is plentiful, volition is valuable.' That's not a self-help line. That's a labor-market argument. If AI makes cognitive outputs cheap, then what's scarce — what you'd expect the market to actually price — is the willingness to do the hard thinking. The friction. The effortful wrestling he keeps describing.
Megan Skiendel: So volition becomes the differentiating asset. Not intelligence — volition.
David Sterling: That's the prediction. And it's testable. If Brooks is right, we should already see wage premiums, or at least hiring structures, rewarding what you might call cognitive friction — the people who argue with the AI output, who won't take the first answer. Are organizations paying for that? Frankly, I don't think they are.
Megan Skiendel: So is he describing something real, or something he hopes will be real?
David Sterling: That's exactly the gap. The principle is compelling — genuinely. But I keep looking for the institutional evidence and it isn't there. What I see is organizations rewarding prompt fluency. 'Can you get the output faster?' That's a different skill than volition. That's — actually, it's almost the opposite. You're optimizing to reduce friction, not preserve it.
Megan Skiendel: Look, I've seen this in rooms. A consultant walks in, runs a problem through their AI stack in forty seconds, produces a clean deck. The client is impressed. Nobody in that meeting is asking 'did you wrestle with this?' They're asking 'is it right?' Volition is invisible if the output looks good.
David Sterling: And that's the structural problem Brooks doesn't quite name directly. Cognitive friction has no price signal. You can't invoice for it. So even if volition is theoretically the scarce resource — the market won't route capital to it until someone can measure it. Which they can't.
Megan Skiendel: Which means institutional pressure runs the other way. The people who insist on wrestling — they're slower. They cost more. They become the friction point, not the value-add.
David Sterling: The principle survives. The incentive structure doesn't support it. And honestly — what the BBC Future research actually found about cognitive offloading, and whether that science even supports the weight Brooks is putting on it, that's the part that makes this whole argument harder to hold together. We'll get there.
Megan Skiendel: But here's what breaks that open — the research the whole argument leans on. BBC Future, May 2026, piece called 'Think outside the bots: How to stop AI from turning your brain to mush.' Reviewing the early work on cognitive offloading. And they put it right there in the piece: 'the science is brand new and we don't have the answers.'
David Sterling: They said that explicitly? In the piece?
Megan Skiendel: Verbatim. And MDPI has a peer-reviewed paper providing academic scaffolding — habitual delegation of memory, reasoning, research to external tools may weaken the capacities being offloaded. May. Hypothesis-stage language.
David Sterling: This is 2011 all over again. Attention-span panic around smartphones. Everyone certain the science was settled. The science was not settled. And yet Brooks's September title — 'AI Is Already Making Us Less Human' — treats cognitive atrophy as structural and ongoing. That's a mismatch. Rhetorical certainty sitting on top of scientific uncertainty.
Megan Skiendel: Okay, but — absence of proof isn't proof of absence. I watched a deputy editor at a place I worked, early 2024, stop writing her own section intros entirely. Handed it to the AI, every week, for four months. And then she had to write one by hand for a legal reason. Couldn't finish it. Sat there. I don't need a peer-reviewed sample size to find that worrying.
David Sterling: That's an anecdote. What does it predict about the next fifty editors who do the same thing?
Megan Skiendel: It predicts — I mean, honestly, that's exactly the question the MDPI paper is trying to ask and can't answer yet. Which is my point. The concern isn't fabricated. The evidence just isn't there to carry the weight Brooks is putting on it.
David Sterling: The cognitive polarization piece is the one that actually survives that scrutiny, I think. The idea — attributed to Brooks in secondary reads, not directly quotable from the primary text, worth flagging — that AI splits people into those who use it to think harder and those who use it to avoid thinking. That's at least a falsifiable prediction. Two diverging populations. Different outcomes.
Megan Skiendel: And the institutional structure almost guarantees which group grows faster. We've established that. The ones avoiding the thinking produce output that looks identical — until it doesn't.
David Sterling: Brooks's two Atlantic essays, June and September, they actually map onto a policy question he never names. Either the intervention is personal discipline, or it's the design of hiring systems, compensation structures, educational institutions that make effortful thinking worth doing. Those are completely different levers.
Megan Skiendel: And nobody is pulling the second one. That's the honest place we've landed, I think. Telling individuals to choose volition — it's real advice, I'm not dismissing it — but once a company standardizes AI into a workflow, the person who insists on wrestling with the problem isn't the value-add anymore. They're the friction point. The institution makes that choice, not the individual.
David Sterling: The question isn't whether AI will make us less human. It's whether anyone is actually designing for the alternative. And I don't see evidence they are.
Megan Skiendel: No. Neither do I. And that's — honestly, that's the unresolved part. Anil Seth's dead sand of silicon isn't the problem. The org chart is.
David Sterling: Good place to stop.
