WebTorrent-based website hosting is back, which means we're due for another round of "decentralization will save the internet" takes followed by the quiet realization that nobody wants to volunteer their laptop's bandwidth for someone else's cat blog. We've been here before.

Peerweb is exactly what it sounds like: host your website across a peer-to-peer network instead of AWS, and visitors load it from other users rather than a central server. It's BitTorrent for web pages, with all the promise and problems that implies. The tech actually works, which is both impressive and irrelevant to whether anyone will use it.

This is the third major swing at decentralized hosting in the past decade, after IPFS's "we're totally going mainstream" moment around 2018 and the blockchain-based hosting wave that peaked with everyone's ETH bags in 2021. Each time, the pitch is identical: break free from Big Tech infrastructure, make hosting censorship-resistant, distribute the costs across users instead of paying CloudFlare or Vercel. Each time, reality shows up with the same invoice.

The problem isn't technical. WebTorrent is legitimately clever engineering. The problem is that decentralization solves a problem most people don't have, while creating a dozen problems they definitely do have. Your website loads slower. Your uptime depends on strangers staying online. Your visitors need to seed content, which means explaining to normal humans why their browser is now also a server. That's a tough sell when Netlify exists and costs literally nothing for most use cases.

Here's what everyone's missing while they geek out over the tech: infrastructure centralized because centralization won. Not through monopolistic scheming, but because it turns out people really, really like it when websites just work all the time, load instantly, and don't require a computer science degree to deploy. AWS didn't conquer the internet by being evil. It conquered by being boring, reliable, and someone else's problem.

The decentralization crowd keeps treating centralization like it's a bug to be fixed, when it's actually a feature users chose. Every survey says people hate Big Tech. Every action reveals they hate unreliable infrastructure more. The gap between stated preferences and revealed preferences in tech has never been wider.

But here's where it gets interesting: the only decentralized hosting that's actually worked at scale is piracy. BitTorrent thrives because the incentive structure is clear—you want the movie, you participate in the swarm. The content itself is the incentive. Peerweb and its predecessors keep trying to replicate that model for legitimate websites, but there's no comparable motivation. Nobody's going to volunteer their bandwidth so some startup's landing page loads faster. The juice isn't worth the squeeze.

The timing here is particularly telling. This launches just as we're watching AI companies strip-mine the open web for training data, and as everyone's suddenly worried about whether their website will even be visible once AI search replaces Google. Decentralized hosting sounds great until you realize you're decentralizing content that LLMs are about to make obsolete anyway. You're solving yesterday's problem with tomorrow's technology while today's problem—AI is fundamentally changing what websites are even for—runs right past you.

The real tell is who gets excited about these projects. It's always developers, crypto enthusiasts, and digital rights activists—people who care deeply about infrastructure philosophy. It's never the people who just want to launch a website for their small business or passion project. That gap between who builds it and who uses it is where these projects go to die.

So will Peerweb work? Sure, technically. Will it matter? Only if someone figures out how to make decentralization more convenient than centralization, which is like trying to make bicycles faster than cars. Possible in theory, fighting physics and economics in practice. The question isn't whether the technology is cool—it obviously is. The question is whether being cool has ever been enough.