It’s All Slop To Me
On the invisibility of epistemic failure
“I never allow myself to have an opinion on anything that I don’t know the other side’s argument better than they do.” — Charlie Munger
In the middle of one of my ever-increasing podcast binges I was listening to one of the many new quasi-famous (which is to say, famous in certain circles) talking heads of the tech industry and he brushed lightly into a niche part of the industry that I happen to know very well. I had been nodding along contentedly, agreeing passively and feeling smart, which is why we listen to podcasts, right? But in this niche that I actually knew, I became alarmed. His tone and confidence didn’t change, even though I’m sure he’s an outsider on this subject. The inaccuracies were subtle. They sounded perfectly reasonable to anyone who didn’t know the area inside and out. But it all added up to make a picture of things that was flat wrong, and wrong in the worst way, because most people wouldn’t know that it’s wrong.
Back in the 1990s, Alan Sokal created an academic uproar by submitting an article to a journal. He was a physicist and was quite unimpressed with some of the newer social sciences that had been peppered with postmodernism. So his test became: can I get an article published in one of these stupid journals if it’s “liberally salted with nonsense”?
And so was born one of the most humbly named journal entries of all time: “Transgressing the Boundaries - Toward a Transformative Hermeneutics of Quantum Gravity”. When it came out as a hoax, the editorial board was embarrassed, sure, but they also asked Sokal to revise and resubmit again (he declined). The journal won an Ig Nobel prize a year later for “eagerly publishing research that they could not understand, that the author said was meaningless, and which claimed that reality does not exist”.
That journal still exists and still publishes, by the way. They’ve never retracted the article.
When Michael Crichton coined the term Gell-Mann Amnesia, named after another physicist, he was talking about newspapers. Those aren’t read much anymore except by a certain segment of ostrich-like Boomers. And so Gell-Mann Amnesia is perceived as a sort of cultural fossil about an older world of media where journalists with little subject depth talked to experts (maybe) and misinterpreted and confidently published anyway and you’d occasionally notice if you saw something about your own expertise.
It didn’t go away, though, it just evolved into something arguably better. Or rather, media is now better, because we get to hear straight from a rotating cast of public intellectuals. These new talking heads are credentialed and titled and often rich. We get to hear their opinions, beliefs, and visions… at least the ones they’ll tell us. These guys have more to say than a beat journalist, and they’re better at saying it too, which is why a one hour podcast on the energy flow through the Strait of Hormuz builds credibility from nothing other than being an hour long.
But that doesn’t mean it’s correct. If one is speaking earnestly, they might be suffering from Localhost Competence. Or simply giving their opinions which, thank God, they are free to do, thank you First Amendment. But don’t ever doubt they’re selling something. It just slid from selling the news to selling their entire worldview.
There are more experts than ever in the public sphere and they’re all competing for your eyeballs and earbuds. The broader they can talk, and the longer, the more their opinions will become your opinions. Our new world is driven by content and influence, but broad knowledge is not deep.
When Google got big, it changed the way we lookup information just as the calculator before it changed how we do arithmetic. AI is changing how we think about knowledge. Broad knowledge is inherently favored because the idea is that an AI can go deep for you anytime you want. So you assemble a bunch of big, broad puzzle piece opinions, often assembled from the worldviews of our favorite public intellectuals, and they’re big and broad so of course they fit together.
We’re all treating depth of knowledge as an on-demand production of a neural network, but the thing the models are really good at is giving you broad, shallow answers. It can give you an easy intro to any topic anytime. A high level summary of Plato’s Phaedo? Sure. A description and prediction of gas prices based on a 20% reduction in crude oil distribution and the impact on tanker size? Yes sir, no problem. How about developing a new transgressive hermeneutics on quantum gravity? Absolutely, Mr. Sokal, here’s some bafflingly reasonable-sounding flatulence.
These new toys of ours are insane. They can summarize and find patterns in huge swaths of information. They can infer and generate and obsequiously support whatever worldview you suggest. It can be bullshit as much as it can be grounded truths and it’s hard to tell the difference.
This is where well-formed expertise can help, but in a different way than it used to. In the newspaper world of Gell-Mann Amnesia, expertise was a method of authority. We’re lazy, so we tune out and trust the expert whenever they sound authoritative, even if they are wrong. But expertise is more important now, because it’s so easy to make confident slop. What an expert provides, what it all keeps coming back to, is verifiability. An expert has the depth to be able to hear something and remark, as George Carlin says, “Well he sounds reasonably intelligent… Aha! he’s full of shit!”
Imagine a pixelated map of the world, like a Minecraft view of the globe. Zoomed out to the top level you can see the continents and the oceans and get a good sense of the overall shape. You can’t use this map to navigate streets in a city, or even see where the main features of each continent are. You just know how the biggest pieces fit together. Broad knowledge gives you this shape and it’s enough to sound authoritative, to reason about the big blocks, but the map is not the territory.
See, you don’t even know what you can’t see at that resolution. If you can’t tell the slop from substance, then it’s all slop to you.
This is not an indictment of the content or the experts. It’s an indictment of me, and if you’re honest, probably you too. The reason we all hate the idea of slop is because we’re terrified that it works on us. Slop is getting better, and smarter sounding slop is even more dangerous and insidious because it can sound so convincing.
The problem with slop is not with the creator, it’s with the audience. Alan Sokal had every right to intentionally publish his fake article, just as every elite has the right to push their worldview. The editors took a look at his credentials and authority, had no idea how to parse and verify what he was saying, did none of the work to peer review, accepted it, and published. Despite claiming expertise, they were an inexpert audience.
Crichton’s original solution to Gell-Mann Amnesia was to disengage and stop consuming media. That seems all but impossible to most of us in 2026, although there are more reasonable steps towards this than most of us care to admit, because we don’t want to admit we’re addicted to the news, or politics, or whatever. I have to remind myself frequently that the world will not end just because so many people (and machines) are so very wrong on the internet. “Strong opinions, loosely held” is a common phrase, except the internet only rewards the first part. Maybe we don’t have to care so much about everything all the time.
But the main thing I’m focused on is the problem of verifiability and maintaining a consistent dose of humility. If I can’t tell slop from substance (and if I don’t know if I can tell, then I can’t tell), then I grab onto that humility and take it to the library. There is always more resolution to add to our pixelated maps. A high resolution in even one or two areas helps us realize when it’s missing elsewhere.
Everyone talks about the permanent underclass that AI will create. Maybe. Maybe it’s less about what the machines do and more about what we do for ourselves. Maybe it’s less about a caste system and more about a way of knowing. Maybe it’s a choice about humility.
