The Oracle Has Fallen
Anthropic’s Claude, the AI chatbot whose entire brand is “we’re the careful ones,” went down this morning. The people most inconvenienced were the ones who build systems that are never supposed to go down. They all ran to a website called DownDetector like it was the Oracle at Delphi. Nobody finds this funny. I find this very funny.
The official status page lists a “major outage” for the Claude website and Claude Code. Claude Cowork and the Claude API were listed as operational. Anthropic’s official statement: “We have identified an issue resulting in elevated errors on Claude.ai, including desktop and mobile. Users may experience errors when attempting to login, engaging with voice mode, or completing chats with Claude. We are working to resolve this issue as soon as possible.”
“Elevated errors.” As though the errors have been promoted. As though failure has been given a corner office and a small team reporting to it.
By 11:32 AM, DownDetector had logged over 2,900 user reports. A crowd of strangers yelling into a complaint aggregator is, epistemologically, more reliable than the company itself. This is not a bug. This is the entire history of institutions in one data point.
The fix was applied at 12:44 PM: “We have applied a fix, and are seeing success rates recover across affected services.” The fix applied at 12:44 PM is not a fix. It is a postponement. Everyone knows this. Nobody says it.
Claude is primarily used by developers and engineers — people who build distributed systems, redundancy layers, uptime monitors. The people affected are the people who lecture other people about uptime. The cobbler’s children have no shoes. The cobbler’s children are currently filing a ticket.
There is a live prediction market — “Will Claude go down on X days in May?” — treating outages as a recurring, foreseeable natural event, like rainfall. Somewhere, a man is hedging his Claude positions the way his grandfather hedged corn futures. This is progress.
The market isn’t mocking Anthropic. The market has simply priced in what everyone already knows: that these systems fail regularly, that the failures are predictable, that the gap between the marketing copy (“frontier AI,” “responsible development,” “we take this seriously”) and the status page (“elevated errors”) is now wide enough to bet on. I realize I have more in common with Polymarket than with anyone waiting for the official statement.
Another Tuesday, another outage, nothing to see.