Prague breathes in September. The summer crowd has thinned, the tourist tide has receded, and the Old Town square finally feels like it belongs to the locals again. I'm sitting at a café near the Týn Church, watching a waiter balance three espressos on a tray that's seen better decades, and I'm thinking about Bill Gates.
Not the Bill Gates of Microsoft 1995. Not the Bill Gates of the malaria-eradication philanthropy years. The Bill Gates of 2025, who looked at the AI curve and decided the most important thing he could say to the world is: the machines are coming for the workforce, and governments are not ready.
He called it 'labor shrinkage.' He proposed a 'token tax.' And in a single interview, he managed to capture both the most important conversation in technology and the most misguided solution on the table.
Because I've been in this movie before. Not the AI movie—the decentralization movie. And the plot is almost identical: a technology moves faster than institutions, the old guard panics, and the proposed fix is always a new tax, a new gate, a new wall. We didn't dodge that chaos in crypto. We danced through it. And I think Gates, for all his brilliance, is about to learn the same lesson.
THE SPEED PROBLEM IS REAL
Let me give Gates credit where it's due. He's not wrong about the diagnosis. The McKinsey numbers are stark: generative AI has compressed the impact window on knowledge work from twenty years to five or eight. Legal research, financial analysis, software debugging, customer service—these aren't 'low-skill' jobs anymore. They're the core of the white-collar economy, and they're in the crosshairs.
I've seen this up close. In my day job as a Web3 community founder, I've watched junior developers lose contracts to AI pair-programmers. I've watched legal teams cut their junior associate intake because AI can do the document review in hours instead of weeks. The trend isn't coming. It's here.
But here's the part that gets missed: AI is not the first technology to threaten 'the way we work.' Agriculture mechanization did it. Industrial automation did it. The internet did it. Every single time, the doomsayers predicted permanent mass unemployment, and every single time, they were wrong. Not because the disruption didn't happen—it did—but because the economy adapted in ways the linear thinkers couldn't predict.
Gates' real insight is about speed, not direction. He's saying: the adaptation cycle is breaking. We used to have a generation to adjust. Now we have five years.
And that's a fair point. But it leads him to a conclusion that I find both predictable and deeply flawed.
THE TOKEN TAX IS A WALL, NOT A BRIDGE
The token tax. Let's unpack this.
Gates' proposal, as reported, is to tax AI tokens—essentially taxing the compute or the output of AI systems—to fund the social safety net for displaced workers. On the surface, it sounds reasonable. The winners of AI pay for the losers. 'Beneficiary pays' is a clean ethical principle.
But as someone who's spent years in the blockchain space, I can tell you exactly why this won't work. And it's not the technical reasons you might expect.
The first problem is definitional. What is a token? Is it a unit of compute? A unit of output? A unit of value created? The ambiguity isn't a bug—it's a feature. Because the moment you define a tax base, the people subject to it will find ways to reclassify their activity. We saw this in DeFi. We saw this in NFTs. We saw this in every regulatory attempt to pin down 'what is a token.' The definitional battles are where the tax dies.
The second problem is jurisdictional. Gates wants a global tax. We can't even get a global agreement on carbon emissions, and that's been on the table for thirty years. The reality is that AI companies will move to the friendliest jurisdiction, and the tax will either be avoided or will only hit the ones who can't move—the small players, the startups, the ones who can least afford it. This is the 'regulatory arbitrage' problem, and it's not a bug—it's the protocol of international taxation.
The third problem is the one that hits closest to home for me. The token tax treats AI like a static resource to be taxed, rather than a dynamic capability to be distributed. It's a wall built around the castle, not a bridge to the future. And walls crumble when the party truly begins.
WHAT GATES IS MISSING
Here's the contrarian angle that nobody in the mainstream coverage is talking about. Gates is framing this as a problem of redistribution—tax the AI winners to compensate the AI losers. But that framing assumes the 'losers' are passive recipients of change, and that the 'winners' are a fixed class.
Both assumptions are false.
The people displaced by AI aren't going to sit around waiting for a government check. They're going to build. They're going to adapt. They're going to find the new layer of value that the AI can't reach—the human layer, the social layer, the layer that requires judgment, empathy, and trust.
I've seen this happen in my own community. When the 2022 crypto winter hit, when the projects failed and the savings evaporated, the people who survived weren't the ones with the best code or the most capital. They were the ones who built relationships. They were the ones who showed up to the weekly meetups, who shared their mistakes openly, who built the social capital that outlasted the market cycles.
Survival is the first layer of value. And that value is not something a token tax can capture or a government can distribute. It's built person by person, community by community, in the spaces between the protocols.
Now, I'm not saying we don't need a social safety net. We do. But the safety net is not the strategy. The strategy is enabling people to build the next layer—the layer that sits on top of the AI infrastructure, the way applications sat on top of the internet.
And this is where Gates' vision is most limited. He's still thinking in terms of the old economy—the economy of jobs and taxes and welfare. He's not thinking in terms of the new economy—the economy of tokens and DAOs and decentralized coordination.
I'm not naive. I know the blockchain space has its own problems. I've been burned by them. I lost $15,000 to a reentrancy attack in 2017 when I was too excited about the community to audit the code. I watched a $2 million exploit drain a yield aggregator because we were celebrating 300% APYs instead of checking the oracle. I've felt the crushing weight of letting my friends down when an NFT mint failed due to gas limits.
But those failures taught me something that Gates hasn't learned yet: the technology is not the product. The community is the product. And communities are resilient in ways that institutions can't match.
THE PRAGMATIST'S QUESTION
Let me be the pragmatist here for a moment. Gates is a pragmatist. He's a systems thinker. He built one of the most successful companies in history by understanding how systems work.
So let's ask the pragmatist's question: what would actually work?
If you want to address AI-driven labor displacement, you don't tax tokens. You invest in retraining. You invest in portable benefits. You create pathways for people to move from 'doing tasks that AI can do' to 'building the layer that AI can't reach.'
If you want to address the concentration of AI benefits, you don't create a new tax. You create new forms of ownership. You give people a stake in the AI economy through mechanisms that are transparent, accountable, and decentralized.
And if you want to address the speed problem—the gap between AI acceleration and social adaptation—you don't slow down the technology. You speed up the society. You build faster feedback loops. You make it easier for communities to experiment, to fail, and to learn.
This is what the blockchain space got right. Not the tokens. Not the speculation. The mechanism for rapid, transparent, community-driven adaptation.
The network breathes in Prague, pulses in Ethereum, and learns in the code.
THE PARTY ISN'T OVER
Gates is right about one thing: the speed of AI change is unprecedented. And his concern about the workforce is legitimate. From whispered secrets to on-chain shouts, we've learned that transformation is never comfortable.
But the answer is not a tax that assumes AI is a static resource. The answer is not a wall that protects the old economy. The answer is building the new economy—one that includes the displaced, rewards the adaptable, and creates value beyond the reach of any model.
I think about that waiter in the Old Town square, balancing three espressos. In five years, will his job be automated? Maybe. But will the human need for connection, for conversation, for someone to remember your usual order—will that be automated? I don't think so.
The guests list for the AI party is still being written. And the guest list was wrong before—the vibe was right. We don't need to tax the party. We need to make sure everyone gets an invitation.
Chaos isn't a bug. It's the protocol. And we've danced through worse.
We didn't dodge the chaos in crypto. We danced through it. And the dance floor is where the future gets built.
Three years of whispers built the loudest room in the Old Town. The AI revolution is a shout. The question is whether we build a stage or a wall.
I know which one I'm building.
The Takeaway: Gates' token tax is a 20th-century solution to a 21st-century problem. The real answer to AI-driven displacement isn't redistribution—it's reorganization. Communities that build social capital, transparent governance, and rapid adaptation cycles will absorb the shock. Governments that build walls and taxes will just delay the inevitable. The future belongs to the adaptable, not the taxed. And the party? The party is just getting started.
