AI, Luddites, and the Iceberg
AI, the knowledge worker reckoning, and a room full of smart people who don’t see it coming.
A few months ago, I was at a party. Not a tech party — a normal party, the kind with lawyers and consultants and finance people. Accomplished professionals. Smart, successful, ambitious, and curious people. Over the course of the evening, I found myself talking about AI with several of them, and I kept noticing the same thing: they’d heard about it. They’d played with ChatGPT once or twice. They found it interesting, maybe a little overhyped.
What they hadn’t done was use it as infrastructure. As the thing that runs underneath everything — thinking, writing, researching, drafting, synthesizing — every single day.
I work in games and artificial intelligence which puts me closer to the tools than most. I recently began advising an AI company that has further pushed me into the directions of vibe coding and working directly with several AI tools.
This means I live inside these tools in a way most people don’t. And standing in that room, surrounded by capable, high-earning professionals going about their evenings, I felt something I couldn’t quite name.
Not superiority — but fear that the average person who does not work in tech is not aware how Ai will change how we work and behave as a society. It reminded me of the Titanic heading towards an iceberg and then sinking while the music played. I have a healthy fear about it but others do not. My fear rises because I am close to the waves of change and I have heard this story before……. in history class in grade school and I see that there may be some similarities with what is to come.
The Luddites Were Not Anti-Technology
In the early 1800s, a group of skilled textile workers began destroying the industrial weaving machines that were replacing them. It only stopped because of harsh laws that were imposed. We call them Luddites, and some people use that word now as an insult — shorthand for people too frightened or stubborn to accept progress. But that reading misses the point almost entirely.
The Luddites weren’t opposed to machines. They were opposed to what the machines were doing to the distribution of wealth. Factory owners were capturing the productivity gains. Workers were absorbing the displacement. The new looms produced more cloth with fewer weavers, but the weavers didn’t share in the surplus they simply became unnecessary.
A similar dynamic played out in the Swing Riots, where agricultural laborers destroyed the threshing machines compressing wages in already-fragile rural economies. The resistance wasn’t philosophical. It was economic and it was desperate.Both movements failed to stop the machines. But they weren’t wrong about the underlying problem: technological revolutions don’t just increase productivity. They redistribute it.
And in the early phases, they almost always redistribute it upward. This leads to societal unrest in most cases and then laws to govern the reactions.
This Time, the Machines Think
Every previous industrial wave automated physical labor. Steam engines replaced muscle. Tractors replaced farmhands. Assembly lines replaced artisans. These transitions were brutal for the people they displaced, but they left a large category of work untouched: the cognitive work. The work that required judgment, language, synthesis, creativity. The work done by lawyers, analysts, writers, consultants, doctors.
AI automates cognition. That’s the thing that’s different this time. Not incrementally different but categorically different. For the first time in history, high-status knowledge work faces scalable automation. And unlike a factory retooling, it scales instantly, at near-zero marginal cost, across every industry simultaneously. According to McKinsey’s 2024 State of AI report, 78% of organizations now use AI in at least one business function — a figure that has risen sharply in just a few years. Generative AI tools have reached hundreds of millions of weekly users globally. The adoption curve inside companies is accelerating faster than any retraining system can follow.
But organizational adoption and individual fluency are not the same thing. That gap — the distance between a company deploying AI tools and an employee actually thinking differently because of them — is where the economic asymmetry is starting to form.
The Compounding You’re Not Accounting For
Here’s what I keep trying to explain to people: the productivity gains from AI don’t add but they compound.
Someone who uses AI to move 20% faster today will redesign their entire workflow in six months. In a year, they’ll have eliminated layers of process that used to require teams. The delta between that person and a peer who never meaningfully adopted these tools isn’t 20% it’s an order of magnitude, and it’s growing.
Industrial revolutions follow a recognizable pattern: technology arrives, productivity rises, labor demand contracts, institutions struggle to adapt, a new equilibrium eventually forms. We are currently somewhere between the productivity surge and the labor contraction.
The adjustment period where the gains haven’t yet been redistributed and the dislocations haven’t yet been addressed is where we’re living now.
In the 1800s, mechanized looms produced more cloth with fewer weavers. Tractors produced more crops with fewer farmhands. Today, AI produces more analysis, more content, more decisions with fewer analysts, writers, and junior professionals. The difference is that this time, the displacement is happening to people who thought they were insulated from it. Junior associates, junior financial analysts, social media marketers, concept artists, the list is ever growing as more tools emerge. This applies to many careers and is hitting juniors who are entering the workforce the hardest.
Capital Will Capture the Gains First
Industrial revolutions have a consistent early-phase pattern: capital wins before labor does. Factory owners captured the gains of mechanized textile production before labor protections existed. Railroad financiers concentrated enormous wealth before any regulatory framework caught up. Platform economies minted a tiny class of tech founders and investors while displacing far larger numbers of workers.
AI will likely follow the same arc. The near-term winners are model builders, infrastructure providers, data owners, and highly AI-leveraged operators people and organizations that can deploy these tools at scale before the rest of the market catches up. If productivity gains flow primarily to that group, wage compression in the broader knowledge economy follows. That’s when the social friction tends to appear. The Luddites didn’t smash looms because they feared change in the abstract. They smashed them because rapid capital consolidation was happening without any social cushioning and no institution was moving fast enough to address it.
There won’t be machine-smashing this time. But there will likely be regulatory pushback, professional guild resistance, legal slowdowns, and cultural skepticism. Some of that will be necessary and legitimate. Governance matters. Safety matters. But suppression has never stopped technological momentum it only reshapes where the value accumulates.
The Real Danger Is Institutional Lag
The most dangerous feature of technological revolutions isn’t the technology itself. It’s the gap between how fast the technology moves and how fast everything else can follow.
Education systems move slowly. Professional accreditation moves slowly. Legal systems move slowly.
Cultural narratives about what work is valuable, what skills are worth acquiring, what a career is supposed to look like — these all move slowly. AI evolves weekly.
This creates very specific risks: skills becoming obsolete faster than retraining programs can respond; credential value eroding as the tasks credentials measure get automated; entry-level roles disappearing before the next generation has a chance to learn through them. Entry-level work isn’t just low-value work it’s how industries have always transmitted knowledge. When those roles vanish, something harder to name vanishes with them.
The Industrial Revolution produced decades of adjustment before child labor reforms, labor protections, and modern employment norms emerged. The disruption wasn’t compressed — it played out across generations.
AI’s disruption won’t look identical, but the economic rhythm is hard to ignore.
AI Literacy Is Becoming Economic Leverage
What I keep coming back to is this: the divide isn’t forming along the lines most people expect. It’s not technical workers versus non-technical workers. It’s not young versus old. It’s not even companies that have AI versus companies that don’t.
It’s people who have genuinely internalized these tools — who think differently because of them, who’ve rebuilt their workflows around them — versus people who treat them as a novelty, a search engine upgrade, something to experiment with occasionally.
The first group moves faster. Produces more. Experiments cheaply. Thinks at a scale that wasn’t previously available to individuals. The second group competes at human speed against AI-augmented peers. That gap is already visible in productivity data. It will become structurally visible in employment data within the next few years.
None of this is destiny. There are genuine choices available — about how to build AI literacy into education, how to redesign work around human judgment rather than human throughput, how to build tools that augment rather than simply replace. The people best positioned to make those choices are the ones who aren’t pretending the disruption isn’t happening.
Back to the Party
I think about that room a lot. Not because those people were foolish — they weren’t. Not because they lacked information — it’s everywhere. But because the transition happening right now has a particular quality that makes it easy to underestimate: it’s invisible until it’s not.
The skilled weavers of 1811 couldn’t see the full shape of what was coming. They saw machines in specific factories, in specific towns. They didn’t see the structural transformation of their entire industry. The dread I felt at that party wasn’t about the people in the room or my inability to talk to them about the upcoming waves and disruptions — it was about recognizing that shape again.
The difference between now and then is that we can see it happening in real time. We have the historical pattern. We have the data. We have, for the first time in any comparable transition, the ability to watch it unfold and make different choices.
The question is whether we will. And I keep thinking: if I could go back to that party, what would I actually say to warn people? And is there anything that we can do to re-route the ship so we aren’t hitting any icebergs. Probably: start now and lets talk more about it.
If you’re thinking about this too — or wrestling with how individuals, companies, and institutions can soften the shock of what’s coming — I’d genuinely like to hear from you. Michael Sorrenti
We’re entering a period where AI literacy, economic policy, education systems, and workplace design will all collide. I don’t think the outcome is predetermined, but I do think the window to shape it is smaller than most people realize.
If you’re working on this problem — whether from the technology side, policy side, education side, or inside an organization trying to adapt — feel free to message me directly. I’m always open to conversations and exchanging ideas.
#ArtificialIntelligence #FutureOfWork #AILiteracy #EconomicChange #TechnologyAndSociety #AITransformation #ProductivityRevolution #KnowledgeWork #AILeadership #AIImpact
