是我一个人感觉到，最近到处都是服务中断吗？（Claude、GitHub、Supabase）
在过去几周，我遇到了多个AI/API的故障——而且不止是一个提供商。
- Anthropic的Claude出现了模型超时和HTTP 500/529错误
- 最近GitHub发生故障时，访问仓库出现问题
- Supabase的数据库/API也出现了小故障
- 甚至主要云服务商也偶尔出现基础设施的短暂故障
单独来看，这些都不算令人震惊。每个系统都有可能出现故障。
但综合来看？感觉故障的频率和影响范围在增加。
这引发了一些令人不安的问题：
- 我们是否为了开发速度而过度优化，牺牲了系统的韧性？
- 小团队是否在不知情的情况下构建在越来越脆弱的技术栈上？
- 这只是可见性偏差，还是故障真的变得更加普遍？
我很好奇这里的其他人是如何看待这个问题的。
查看原文
Over the past few weeks I’ve hit multiple AI/API outages — and not just one provider.<p>Model timeouts and HTTP 500/529 errors from Anthropic’s Claude<p>Repository access issues during a recent GitHub disruption<p>Database/API hiccups from Supabase<p>Even intermittent infra blips from major cloud vendors<p>Individually, none of these are shocking. Every system fails.<p>But collectively? It feels like the frequency and blast radius are increasing.<p>It raises some uncomfortable questions:<p>Are we over-optimizing for speed of development at the cost of resilience?<p>Are small teams unknowingly building on increasingly fragile stacks?<p>Is this just visibility bias — or are outages genuinely becoming more common?<p>Curious how others here are thinking about it.