← Back to Insights

The Cult of "Most"

A large, loud, unmeasured claim shown as an oversized megaphone with expanding sound rings and no scale, next to a small tightly connected cluster of real data points that does have one

This one's a rant. Fair warning up front, because I use AI for a lot of my work, it helps me structure my thinking, tidies up my spelling, catches the sentences that only make sense inside my own head, and I'll happily say so in public. But there's a trend I keep tripping over on LinkedIn, and it's been building for months, and I've finally reached the point where I need to get it out of my system rather than mutter at my phone every time it happens.

It's the "Most" posts.

You know the ones. Most companies are doing AI wrong. Most people are using ChatGPT completely backwards. Most leaders don't understand what's coming. Every second post in my feed seems to open with some version of the word "most," swung around like a full stop that's already won the argument before the argument has started.

Where's this coming from, exactly?

Here's my actual question, and it's not rhetorical, I mean it literally. How do you know? Where's the research? Where's the survey, the sample size, the methodology, anything at all that would let me check your working? I am, for the record, never doing the thing they've just told me most people are doing. Not once have I read one of these posts and thought "ah yes, that's me, bang to rights." So either I am a statistically improbable outlier every single time, which would be its own strange kind of achievement, or "most" was never a real number to begin with. It was a vibe wearing a number's clothes.

Either I'm a statistically improbable outlier every single time, or "most" was never a real number to begin with.

And look, I wouldn't mind nearly as much if people just told the truth about the size of their sample. "Most companies I'm helping right now are doing this wrong." "Most people I know personally are using AI this way." Fine. Grand. That's a claim I can actually weigh, because it comes with a scope attached, your client list, your mates, your LinkedIn bubble, whatever it is. I might even agree with you. But that's not what gets written. What gets written is the universal, unscoped, unfalsifiable "most," dressed up to sound like it's speaking for the entire market, when what it's actually speaking for is one bloke's inbox.

Maybe it's just my tic

I'll own that some of this might be me. I've got a bit of a thing about absolutes and relatives in language, always have, and I notice the gap between them faster than most people would bother to. Say "some," "several," "a few clients of mine," and I'll nod along happily. Say "most" or "everyone" or "nobody," and some small pedantic gremlin in my head sits up and demands to see the workings. That's a me-problem as much as it is a them-problem, and I'm happy to admit it.

But I don't think it's only me, because the pattern is too consistent to be one grumpy Australian's pet peeve. It shows up post after post, account after account, always with the same shape: a sweeping claim about a hypothetical majority, zero evidence, and a comment section full of people agreeing enthusiastically with a number nobody actually has.

This is what AI slop actually looks like

I've ranted about the phrase AI slop before, and back then I was mostly talking about the surface texture, the six-fingered stock photo, the paragraph that reads like it was extruded rather than written. This is the same disease, just wearing a different coat. It's not the prose that's slop here, plenty of these posts read perfectly well. It's the confidence. It's the certainty about a "most" that was never measured, delivered in a tone that dares you to disagree, because who's going to argue with "most companies," it sounds so solid, so researched, so obviously true that questioning it feels like you're the one being difficult.

The lazy version

"Most companies are doing AI wrong." No source, no scope, no sample. Just a big confident number that was never a number.

The honest version

"Most of the companies I'm working with right now are doing AI wrong." Same sentence, one clause added, and suddenly it's a real, checkable, personal claim instead of a made-up statistic.

That one clause costs you nothing. It doesn't make the post less punchy, if anything it makes it more credible, because now I know you're talking about something you've actually seen rather than something you've decided must be true because it makes a tidy hook. But adding it means admitting your evidence is smaller than "most," and I think that's the actual reason nobody does it. The unscoped absolute feels more important. It sounds like you've cracked the market instead of just noticed a pattern in your own client list, which, let's be honest, is all any of us are really doing most of the time.

So no, I'm not going to write the post everyone else writes. I'm not going to tell you most people are doing anything, wrong or otherwise, because I genuinely don't know, and neither, I'd wager, do most of the people posting about it. What I do know is what I'm seeing, in my own little corner of this, on my own clients, in my own feed, and that's a "some," not a "most." It's smaller, it's honest, and it's the only version of the claim I can actually stand behind.

More rants, projects and the occasional useful idea over on the blog, or head back to the Apps & Games page if you want something less grumpy.

Reactions & comments

Tap a reaction, or leave a comment below — sign in with GitHub. Keep it kind and constructive.