Watch the video first.
University of Central Florida, graduation day, May 2026. The invited speaker is the woman at the podium, a real estate executive, that decided to speak about why AI is the next Industrial Revolution. What happens after this is not what she expected. Not scattered dissent. Something collective and immediate, like a room deciding something all at once.
She stops mid-sentence. Turns around. Her hands go up — not in triumph, in something between confusion and surrender. She tries again. Same wall of sound.
I’ve watched it many times now. And each time what gets me isn’t the booing. It’s the half-second before she turns. The moment you can see she didn’t see it coming.
That’s the thing worth thinking about.
Not the booing itself. The gap in perception.
Here’s someone who prepared a speech, was invited to give it, presumably talked to people beforehand about what to say. And nobody — not her, not whoever organised the event — thought: maybe don’t open with this particular line, to this particular audience, right now.
That’s not carelessness. That’s a worldview. One in which the story of AI-as-progress is so obviously true, so widely accepted by the people one talks to, that it wouldn’t occur to anyone to question whether the room would receive it differently.
The room received it differently.
A few months before Florida, Kristalina Georgieva stood in Davos and used the word tsunami.
She’s not a “radical”. She’s the managing director of the IMF — an institution whose job is to be sober about these things. And sober is exactly what she was. Sixty percent of jobs in advanced economies affected by AI, she said. Entry-level roles hit first. The ones where people have always started — the jobs that turn someone with a degree into someone who actually knows something through the slow accumulation of doing things badly and then less badly and eventually competently. Those jobs.
She also said the middle class would be “inevitably” affected.
Inevitably.
That word does a lot of work. It forecloses something. It says: this is not a policy choice, this is not a negotiation, this is the weather. Get an umbrella.
The students in Florida weren’t in Davos. They didn’t need to be. They’d been watching what happened to people a few years ahead of them. The job postings that used to be there and aren’t. The internships that got automated. The entry points that closed while everyone was having the debate about whether they’d close. A Harvard Kennedy School poll from 2025 found that most — not a vocal minority, not the anxious ones, most — recent American college graduates think AI is a threat to their futures.
The booing was the real time technicolor translation of that number. Not the language of polite disagreement Just a web of angry sounds.
There were a lot of booing moments through human history, so why does this moment feel different to me?
Young people and new technology have always gotten along. Almost a social law — the young adopt faster, find the opportunity before the anxiety settles in, end up positioned better because they didn’t hesitate. Electricity. Computing. The internet. Smartphones. The pattern holds across every wave.
Except here’s what I’m watching: the pattern is not holding.
The people in that arena aren’t Luddites. They’re not afraid of technology in some general way. They use AI tools constantly — for their work, their creative projects, their daily lives. They grew up on platforms built by the same people who are now building the systems replacing their jobs. They understand, probably better than most people twice their age, exactly what this technology can do.
What they’re rejecting isn’t the technology. It’s the story. The one where displacement is just disruption and disruption is just progress and progress is always worth it and anyone who says otherwise is scared of the future.
They’ve heard that story. They’ve heard it from people who will be fine regardless of how this goes.
They are not buying it.
Varoufakis uses the word technolords. Worth sitting with.
These people are not only powerful, because that very power operates like ownership — ownership of the narrative, of what counts as progress, of who gets to define the terms of the debate. Politicians get called out of touch all the time. That’s normal now, almost background noise. But watch what’s happening and you start to wonder if that complaint is migrating. Attaching itself to a different group. One that perhaps assumed, because it wasn’t elected, that it was exempt.
That assumption may be getting tested.
The jobs aren’t the whole story, either.
Virginia, and other states — young people showing up to protest data centre expansion. Not abstractly. Specifically. The water consumption. The electricity. What these facilities do to local grids, to communities that suddenly have a server farm next door running twenty-four hours a day. The Packard Foundation published something this year connecting this activism to a broader pattern — AI infrastructure becoming a political flashpoint, climate anxiety and technology anxiety fusing into something that doesn’t have a clean name yet.
There’s a concept worth naming: the Jevons Paradox. The observation that more efficient systems don’t reduce consumption, they expand it — because efficiency lowers the cost of doing more, so more gets done. Applied here: more efficient AI doesn’t mean less energy. It means more AI. The generation that grew up being told climate was the crisis that would define their lives is watching this and being told it’s progress.
And then the schools. A study from earlier this year, eight secondary schools in the UK. Sexualised AI-generated images of students circulating through peer networks. Almost always girls. Teachers and students both without the language, the legal framework, the institutional support to respond. One in thirteen British teenagers had been through this. The researchers found that students and teachers connected it, consistently and without being prompted, to a broader rise in misogyny — normalised and amplified by the same algorithmic architecture that the same industry built and continues to profit from.
Jobs. Climate. Sexual violence made easier by technology.
Not three separate complaints. The same complaint arriving from different directions. What’s being built is being built at their expense. The people building it are — at best — not thinking carefully about that. At worst, they know and are proceeding anyway.
I want to be careful about over-reading a video.
Florida was a moment. Virginia was a protest. The petitions against mandatory AI tools in universities are real but scattered. There’s no movement yet in any organised sense. No programme. No political vehicle. What there is — and this might be the more important thing — is a feeling that is looking for a form.
Sometimes feelings that look for forms find them.
What seems genuinely new is the scope. AI anxiety used to be narrow — jobs, mostly, some privacy concern. What you can see now, in the range of things that are generating pushback, is something wider. The labour market worry and the climate worry and the deepfake worry are starting to feel like one thing rather than three. Expressions of a single problem: the system makes decisions affecting everyone and is accountable to very few, and the few it’s accountable to aren’t the ones absorbing the costs.
That kind of realisation — when it takes hold — is hard to co-opt. You can co-opt a specific demand. You can respond to a call for a regulation, a policy change, a feature. You can’t easily respond to a call for a different story about who matters. That requires something the people telling the current story are structurally disinclined to provide.
The technolords have always been good at moving fast enough that the critique becomes obsolete before it can organise. That works when the critique is specific. When the critique is of an arrangement — of who decides and who absorbs the consequences — it’s a different problem.
She really didn’t see it coming.
What’s striking here is not just the booing. It’s the assumption underneath it — that this room, these people, would receive this message the way it’s been received in boardrooms and conference halls and TED stages for the past several years.
The compact that made that assumption reasonable for most of the modern era was something like: yes, disruption is real, yes it’s uncomfortable, but broadly, over time, the gains get distributed widely enough that it’s worth tolerating. People believed this because often enough it was true. Or true enough to sustain the belief.
Right now, to a lot of people — particularly young ones — it is not feeling true.
When a story stops feeling true, the people who keep telling it start to look like something other than visionaries. They start to look like someone who prepared the wrong speech for the wrong room.
She threw her hands up.
At some point that gesture stops being some kind of question mark and starts being something else. I’m not sure we’re there yet. But it feels like we’re closer than most of the people at those podiums understand.
Brilliant...the tech-narrative will hit a hard physical wall: energy dissipation. Right now, AI operates as an open-ended sink of resources, externalizing its massive thermodynamic costs—electricity and water—onto the public while centralizing the returns. The turning point of this social friction won't be philosophical; it will be energetic. If AI fails to rapidly trigger a deflationary energy revolution—delivering cheap, abundant, clean power to the masses—the narrative collapses entirely. Without that energetic payback, the "technolords" are just accelerating Jevons' paradox, burning collective resources to automate human labor. Energy data, not consensus, will dictate the final outcome.