March 17, 2026
Vibes.exe has stopped responding
Claude Is Having an Outage
Panic, memes, and a “works for me” guy as users ride the Claude wobble
TLDR: Some users say Claude’s Cowork tool is glitching and locking them out, while others report no issues at all. The comments split into memes and mild panic versus cool “works for me” takes, sparking a bigger debate about whether fast-moving AI tools should feel rock-solid or just good enough right now.
Cue the chaos: users say parts of Claude are wobbling, with the Cowork tool allegedly bricking itself when it hits errors—basically, once it glitches, you can’t get your session back. The original post praises Anthropic for fast, impressive releases, but calls these bugs “disconcerting,” comparing them to seeing hospital equipment misbehave. It’s not life-or-death software, sure, but the vibe is: shouldn’t this stuff feel sturdier?
The comments turned it into a spectacle. One jokester declared, “I’m a viber… what do I do now?” while another threw in faux doom with “Time to throw in the towel!” A third went full drama: “jobless rn.” Then the plot twist: a calm voice chirped, “Claude’s been working fine for me,” splitting the room between panic-posters and the “works-for-me” faction. Even a meta-quipped, “guess that’s why there aren’t many comments here,” as folks waited for the AI dust to settle.
The fight isn’t just about a glitch—it’s about expectations. Some want hospital-grade reliability, others accept “move fast and break things” as the cost of progress. Meanwhile, people watch the flickery code interface like it’s a haunted TV, muttering, “this is a solved problem,” yet admitting the tool still gets the job done. Drama? Absolutely. End of the world? Not today.
Key Points
- •The article reports that Anthropic’s Cowork can corrupt its session when API errors occur, making the state unrecoverable.
- •The piece highlights concerns about reliability in AI software when such bugs appear.
- •Cowork is described as non–mission-critical, with failures unlikely to risk human safety.
- •The author observed flickering in the Claude Code TUI, though it remains functional.
- •The article reflects on the challenge of balancing quality with rapid feature delivery in AI development.