# Five Days Between Warning and Proof By [jonathancolton.eth](https://paragraph.com/@jonathancolton.eth) · 2025-11-18 --- Five days ago, Vitalik Buterin and the Ethereum Foundation published a warning: "If AWS, GCP, and Cloudflare went dark, most apps would too." Today, Cloudflare went dark. X stopped loading. ChatGPT threw errors. Spotify went silent. McDonald's kiosks displayed "service unavailable." Even DownDetector—the site you check when things break—was down. Five days between warning and proof. This wasn't bad timing. It was validation. And it shows exactly why the Trustless Manifesto matters—and why good intentions aren't enough to stop centralization. The Trap We Keep Falling Into ----------------------------- Here's what happened this morning: Cloudflare, a company that protects websites from attacks and keeps them running during heavy traffic, experienced a "spike in unusual traffic" at 11:20 UTC. Within minutes, roughly 20% of the internet went down. Nobody forced these companies to depend on Cloudflare. The protocol is open—you can run your own infrastructure. But Cloudflare made centralization so convenient that it became the default. And defaults become dependencies. Dependencies become single points of failure. This is how centralization actually works. Not through force or capture, but through making alternatives feel unreasonable. Think about email. Right now, you could technically run your own mail server. The protocol is still open. But Gmail's spam filters and reputation systems have made self-hosting so painful that most people don't even consider it anymore. They genuinely prefer Gmail—not because they're forced to, but because it's easier. Until Gmail goes down. Then everyone wishes they had alternatives. Crypto is heading down the same path. Ethereum's protocol is permissionless. You can run your own node. But when Infura and Alchemy handle the majority of requests, we've built the same fragility that just took down half the internet. Why Good Principles Aren't Enough --------------------------------- The Trustless Manifesto lays out three laws to prevent centralization: * No critical secrets * No indispensable intermediaries * No unverifiable outcomes These are necessary. They prevent the most direct forms of control. But they don't address why people choose centralized systems even when alternatives exist. Between outages, centralized convenience always wins. Today's Cloudflare failure reminds everyone why decentralization matters. But by tomorrow, that memory will fade. Next week, new developers will default to hosted RPCs because they "just work." The drift happens in three stages, and the manifesto only addresses the first: **Stage one** is about what you _can_ do. Cloudflare doesn't prevent you from running your own infrastructure. The technical options remain open. **Stage two** is about what seems _reasonable_. When everyone uses Cloudflare, running your own infrastructure starts to feel extreme. Paranoid. Why would you do that when this works fine? **Stage three** is about what people _want_. Eventually, the preferences shift. People don't just tolerate centralization—they genuinely prefer it. They're not coerced. They choose convenience. This is where drift becomes permanent. Technical safeguards don't matter if nobody wants to use them. What Actually Works ------------------- Solving this requires working on three levels at once—not just the technical layer, the manifesto focuses on. **First: Make trustless systems actually usable.** The manifesto's three laws are right. But technical openness means nothing if ordinary people can't participate. Make running your own node genuinely accessible—not "technically possible if you read documentation for six hours." **Second: Change the culture.** Between outages, you need to make decentralization feel normal, not extreme. Use today's failure. When advocating for trustless infrastructure, say: "Remember November 18th, 2025, when Cloudflare took down 20% of the internet? That's what happens when we depend on a handful of providers." Track "temporary" centralization that becomes permanent. When a rollup launches with a centralized sequencer and a "decentralization roadmap," watch it publicly. Celebrate when they ship. Call it out when they don't. Make running your own infrastructure heroic, not paranoid. When someone self-hosts, amplify it. **Third: Make trustlessness profitable.** The manifesto says it plainly: "A protocol that relies on unpaid altruism will decay." Make neutrality the winning economic strategy, not just the virtuous one. Reward people who maintain neutral infrastructure without giving them power to extract rents. Make it more profitable to run trustless systems than to depend on centralized providers. When sovereignty pays better than convenience, drift becomes reversible. The Real Test ------------- The manifesto will succeed or fail based on whether it changes what people _want_, not just what they build. Success means developers want to build trustless systems, users prefer self-sovereignty, and the market rewards decentralization. Failure means everyone signs the manifesto, agrees with the principles, then builds centralized systems anyway because "users want UX." We wait for the next outage to remind us why this matters. Today reminded us. But reminders fade. By next week, most people will go back to trusting Cloudflare because convenience wins every time—until it catastrophically doesn't. We shouldn't need a crisis to trigger change. That's the whole point of building trustless systems now. So they're ready when we need them, not scrambling to deploy alternatives while everything burns. Why I Signed ------------ I signed the Trustless Manifesto while writing this. Five days between their warning and today's proof—that's how fast theory becomes reality. The manifesto names the problem correctly. What I'm adding is why it keeps happening: convenience beats principle until crisis hits, and drift consolidates through shaping what people want, not just controlling what they can do. Understanding the mechanism matters because it tells us what actually works. Technical safeguards plus cultural change plus economic incentives. Not principles alone. Trustlessness is expensive—constant vigilance, ongoing work, choices that feel wasteful until they save you. But centralized convenience is more expensive. You just pay the price later. When your government bans your coordination tools. When your platform decides you violated the terms of service. When Cloudflare goes down and takes half the ecosystem with it. The internet paid a small price today—a few hours of disruption. But it's a reminder that centralized infrastructure always fails eventually. The only question is whether we'll have alternatives ready when crypto's moment comes. Sign the manifesto. But more importantly, understand why drift happens, and build systems that make trustlessness attractive before we desperately need it. Because the only sustainable defense is making freedom more attractive than control—and more accessible than crisis. * * * **The Trustless Manifesto: November 13, 2025** **Cloudflare outage: November 18, 2025** **Five days between warning and proof.** [Sign the Trustless](https://trustlessness.eth.limo/general/2025/11/11/the-trustless-manifesto.html) --- *Originally published on [jonathancolton.eth](https://paragraph.com/@jonathancolton.eth/five-days-between-warning-and-proof)*