There’s a widening gap in how people understand AI at the moment, and it’s got very little to do with intelligence in the usual sense. It’s not about who’s clever and who isn’t. It’s about when you came across it, how you used it, and—crucially—which version of it you actually experienced.
Spend a few minutes scrolling and you can almost sort people into loose categories.
A large chunk had a go on ChatGPT sometime last year, usually the free version. They asked it a few questions, poked at it, maybe tried to trip it up a bit. It gave some decent answers, then something slightly off, then something completely wrong. A hallucination here, a strange phrasing there. Enough to raise an eyebrow.
And, that was that.
They;ve moved on, but the impression stuck. “Clever, but not reliable.” It settled into place as a kind of mental bookmark—filed under interesting experiment. Not revisited, not really questioned.
You still see the aftershocks of that now. Clips doing the rounds of voice assistants fumbling basic questions, people laughing at odd replies, presenting those moments as if they’re definitive. And to be fair, those clips aren’t fabricated. They happened. They’re real interactions.
But they’re also narrow. They tend to sit at the edges—older models, constrained modes, or scenarios designed to provoke failure. What’s happening is that a thin slice of the system is being treated as the whole thing.
It isn’t.
There’s a lag here. A quiet one, but significant. People are reacting to different versions of the same technology, often without realising it.
Because underneath all of this, things have moved on. Quite a bit, actually. The models are more stable. They hold context better. They reason more consistently. You can move between text, images, even voice, without it feeling like a party trick. It’s less about “look what it can do” and more about “this just works now”—which, ironically, makes it less visible.
And then there’s access. That’s the part people tend to overlook. The free version, a paid tier, an API plugged into someone’s workflow—those aren’t just different entry points. They’re genuinely different experiences. Same label on the tin, different machinery inside. I suppose a bit like a local supermarket, maxi super market and the Whole seller.
So you end up with this odd situation where two people say they’ve “used AI”, and they might as well be talking about different tools. One remembers something patchy and unpredictable. The other is using it daily—writing, coding, organising, thinking through problems—without making a fuss about it.
No big reveal. Just gradual absorption.
That’s where the gap opens up.
We’ve seen versions of this before. Early internet, early smartphones—there’s always that stretch where public perception lags behind what the technology can actually do. But this feels quicker, more compressed. Updates aren’t annual anymore; they’re constant. Quiet upgrades that don’t announce themselves unless you’re paying attention.
So the version people dismissed six or twelve months ago isn’t quite the one that exists now. But unless something forces a re-evaluation, that original judgement just… lingers.
What’s being debated online, then, often isn’t the current state of AI. It’s a kind of echo—older experiences, clipped failures, second-hand impressions. A moving target, reduced to a fixed idea.
The thing people think they’re arguing about has already shifted slightly out of reach. or sometimes massively out of reach.
And that’s the uncomfortable bit. Not that anyone’s entirely wrong—but that they’re right about something that’s no longer fully there.
If anything, the gap’s still widening. Not closing. Nut like all emerent tch its developing with human limitations nd bias, rate than skilled output.