That’s why so many people think AI is a bubble right now. Not because it doesn’t work. Not because it won’t change things. But because the industry stopped the show and passed the hat before most people had felt the pull of the next beat.
When I was younger, just starting out as an actor, I was looking for training and work that would broaden my range. I was classically trained in grad school, and I supplemented that training with improv classes, writing classes, and the like. One of my favorite forms of work and training was at Renaissance fairs. As a mud beggar, I performed with my cohorts in a mud pit with a deep end you could disappear into and come up behind the audience if you worked it carefully. It was three of the greatest summers of my life.
As an actor, I always think about what the character wants and how they are going to get it. As a mud beggar, my want was clear: money for food and drink. That was true when I was doing street work, and it was true when the troupe was doing a mud show at the pit. Make money. Get the hat full.
On the street, I did that by doing funny bits and involving the audience. Make the crowd laugh. Pick on someone for laughs. Coming up with clever bits allowed me to fill my hat. In the mud shows, we structured the performance to build to a moment where the next beat was going to be hugely funny or reveal something great, and then we stopped. At that stop, someone would ask the audience if they wanted to see what happened next, and if they said yes, which they always did, we’d pass the hat.
We developed what’s called a good hat line. A funny show, paused at just the right moment, usually led to a good hat.
That pause, the moment before the hat comes out, is the bubble.
It’s not the trick. It’s not the ending. It’s the suspended second where the audience decides whether they care enough to pay to see what happens next. When it works, the hat fills. When it doesn’t, it doesn’t matter how good the ending is. You already lost the moment.
That’s why so many people think AI is a bubble right now. Not because it doesn’t work. Not because it won’t change things. But because the industry stopped the show and passed the hat before most people had felt the pull of the next beat. We explained the trick. We promised the ending. We talked about inevitability. We talked about replacement. We talked about valuation. All before the audience had leaned forward.
If the AI bubble breaks, this is what that actually looks like. The money thins out. The loudest promises stop working. The crowd gets smaller and more skeptical. What’s left are the performers who can still earn a hat from people who already know the trick. That’s not collapse. That’s discipline.
And if it doesn’t break, if the hat still fills, then it means something else entirely. It means AI crossed the threshold from spectacle to craft. From something you’re told will matter to something you’ve already decided you want more of. Fewer demos. Less noise. More moments where someone says, wait, do that again.
The bubble isn’t the explosion everyone’s waiting for. It’s the pause. And right now, the industry is finding out whether anyone actually wants to see what happens next.
The Pause
Sometimes the show felt loose or slightly off, and the hat was great. That took a while to make sense of. After a while, you started paying attention to where things actually shifted. Not at the big laugh. Not at the stunt. Earlier than that. There was a brief stop in the show. Just long enough for the audience to notice something had changed.
That moment didn’t last long. You could feel it when it arrived, and you could feel it when it slipped past without anyone quite noticing. The show kept moving either way. People laughed. They stayed. The difference showed up later.
The hardest shows weren’t the bad ones. They were the familiar ones. The audience had seen the bit earlier in the day. They knew the rhythm. They knew what came next. Everything still worked. People laughed at the right places. It just didn’t work in the same way.
Out of sequence
The conversation around AI has a similar feeling right now. Not the technology itself. The timing of how it’s being presented. There’s a lot of stopping happening. Big pauses. Big claims about what comes next. But for a lot of people, the moment hasn’t quite arrived yet. Instead of feeling pulled forward, a lot of people are being asked to take it on faith. They’re being told how powerful it is. How inevitable it is. How much it will change things.
That creates a strange distance. The pause is there, but the curiosity isn’t always with it. When that happens, the pause doesn’t feel like anticipation. It feels like pressure. Pressure is often what people mean when they reach for the word bubble, even if they’re not quite saying it yet.
What people think is happening
A lot of the debate about an AI bubble assumes the problem is hype or valuation. Too much money. Too many promises. Too fast. That explanation is neat. It fits familiar stories about tech cycles. It gives everyone something concrete to point at. It also keeps the focus on the surface of things, where the numbers are loud and easy to argue about. What it doesn’t spend much time on is how the moment actually feels to the people being asked to care. Whether they’ve actually experienced anything that makes them want to see what comes next. Whether the pause is happening because curiosity is peaking, or because the story is being forced to stop too early. That distinction tends to get lost when the conversation stays focused on money instead of momentum. It’s easier to argue about valuation than to admit that the audience might not be leaning forward yet.
That’s when the word bubble starts doing a different kind of work. It becomes a way to name discomfort without having to talk about timing. A way to suggest something is wrong without asking whether the moment has actually been earned. It lets people argue about collapse instead of sitting with the quieter possibility that things are simply out of sequence. That pause, when it shows up this way, isn’t a peak. It’s a mismatch. And mismatches don’t explode. They either get corrected, or they quietly thin out the room. The question isn’t whether AI survives a bubble, but whether it ever reaches a moment people are genuinely waiting for.
If it breaks
If the AI bubble breaks, it probably won’t look like an explosion. It’ll look quieter than that.
Less drama. More thinning. Funding tends to pull back before anyone announces it. The loudest tools get quieter, not because they stop working, but because the promises stop escalating. Attention shifts away from spectacle and toward whatever still earns it without explanation. What disappears first isn’t usefulness. It’s tolerance. People stop indulging tools that need to be explained, justified, or defended before they can be felt. The room gets quieter, but it also gets more honest. What remains are the tools that don’t need a pause built around them. They’re already inside the work, already doing something small and specific that people would miss if it went away. Nobody argues for them. They just keep getting used.
If it doesn’t
If the bubble doesn’t break, that doesn’t mean the tension was imaginary. It means the pause finally lined up with curiosity. Enough people decided they wanted to see what came next. In that case, AI stops needing to be defended or explained. It moves out of the spotlight and into the background, where tools usually end up once they’ve earned their place. The show doesn’t stop anymore because it doesn’t have to. There’s no need to announce what’s coming next. People are already staying because they want to see it. At that point, the hat doesn’t come out as a test. It comes out as part of the rhythm. Not because anyone was convinced, but because the moment arrived when it was supposed to.
Who the pause is for
One thing that gets blurry in the AI conversation is who the pause is actually aimed at. Investors are listening for one signal. Builders are listening for another. The people expected to use these tools are often somewhere else entirely.
In the mud pit, that part was never abstract. You knew who you were asking because they were standing right there. You could see when attention shifted. You could feel when it didn’t.
Passing the hat to the wrong crowd wasn’t a theory. It showed up immediately. A lot of AI messaging assumes that if you stop the show loudly enough, everyone will stop with you.
But attention doesn’t move as a group. It splinters. Some people lean in early. Some drift away. Most just keep going until something actually interrupts their work.
The takeaway
The AI bubble isn’t a verdict waiting to be delivered. It’s a moment waiting to line up. A pause that only works if it arrives when someone is already leaning forward.
In the mud pit, you learn quickly that you don’t get to declare that moment. You can’t explain your way into it. You can only notice when it’s there, or notice when it’s already passed.
At ThoughtLab, that’s how we look at technology shifts like this, not as inevitabilities to announce or futures to scare people into, but as moments that have to be earned with the right audience, at the right time. If the hat feels light, it’s not because the ending isn’t impressive. It’s because the pause didn’t meet the people it was meant for.
