Sonnet 4.6 Elevated Rate of Errors

Claude’s Sonnet 4.6 hit a late‑night wobble, throwing extra errors from about 11 pm to 1:50 am Pacific. It’s fixed now, says the official status page. But the real show was the comments, where users turned a routine hiccup into a roast. The harshest zinger: it’s “probably more newsworthy when it’s up, not down,” a burn that had the crowd cackling. Others demanded Anthropic open the checkbook, pointing to talk of a “$30B run rate” and yelling: if the cash is flowing, the servers shouldn’t be wobbling.

Then came the literary shade: a “Flowers for Algernon moment,” implying the smarter the bot gets, the more it stumbles—brutal. Some tried to downshift the drama, noting Hacker News loves a front-page panic but that models “are up or down on any given day,” linking back to the status page receipts. Meanwhile, another jab mocked the company’s big-brain aura—“mythos” won’t fix outages—while devs groaned that everything from the website to the API and coding tools took a hit. The vibe split cleanly: pragmatists saying “it was short and overnight, chill,” versus uptime purists saying reliability is the whole product. Verdict from the crowd? Fix landed, but trust took a dent—and they want fewer sonnets, more stability.