If you felt a strange disturbance in the force yesterday — like thousands of startup founders suddenly cried out in terror and were silenced — that was Claude going down.

Anthropic's popular AI chatbot experienced an outage that left its most devoted users staring blankly at their screens, suddenly confronted with the horrifying reality of having to think for themselves. And nowhere was the collective meltdown more palpable than right here in San Francisco, where Claude isn't just a tool — it's a coworker, a therapist, and for some people, probably the closest thing to a best friend.

"There goes my day," one SF resident lamented, capturing the sentiment of a city that has apparently outsourced basic cognitive function to a language model.

Look, we get it. Claude is genuinely useful. It writes code, drafts emails, summarizes documents, and does roughly 40% of the actual work at any given SoMa startup (conservative estimate). But the sheer panic that erupted over a temporary service interruption should give us pause.

We've spent the last two years listening to San Francisco's tech elite lecture the rest of the country about AI safety and existential risk. But the real existential risk, apparently, is Claude being unavailable for a few hours on a Tuesday. The dependency is real, and it's a little embarrassing.

Here's the fiscally conservative take nobody asked for: if your entire business grinds to a halt because one chatbot goes offline, you don't have a business — you have a subscription. Maybe it's worth thinking about what happens when the tool you've built your workflow around has a bad day. Or gets more expensive. Or changes its terms of service.

Anthropic got things back online relatively quickly, and the city's productivity presumably resumed. But the episode was a useful little stress test — not of the AI, but of us. And we did not pass with flying colors.

Maybe keep a notebook handy. Just in case.