Holy shit, Anthropic just cut Claude API limits by 60% overnight. No warning. No email. No announcement. Your production app that was working yesterday? Now it's throwing rate limit errors every 30 seconds. This is why you never bet your entire stack on someone else's API.

Overnight limit cuts with no warning are exactly why single-provider production apps are risky. FastRouter's automatic failover means when Claude hits a rate limit your requests reroute to the next available provider in under 10ms. Nothing to configure when the limit drops, it just keeps working.