Dev questions Anthropic ops: Claude Code source leaked to prod, backbone outages 'routine'
cuenca · x · 2026-10-29
Developer @cuenca revisits Anthropic's recent incidents: a Claude Code build with source code was accidentally pushed to production, and the Claude Code lead's postmortem conclusion was, surprisingly, to "automate MORE." He also asks how often Anthropic's backbone (LZB) goes down this year—the tone of official communications suggests it's routine. He ties this to a broader observation about modern software security: the theoretical ten layers of defense are rarely stress-tested because no single layer carries full load, so none get seriously hardened.
More from Companies & People
- The real question behind Mithril: can one person plus AI run an entire drug pipeline? — FinanceYF5 · 2026-10-29
- 'Pipeline-in-a-Person': one person orchestrating AI and experts across multiple drug pipelines — FinanceYF5 · 2026-10-29
- 20VC: Anthropic files S-1 for $2T IPO, Microsoft lays off 27,000, Meta raises $95B — 20VC · 2026-10-29
- Microsoft's 4,800-person layoffs save what its AI data center spend burns in two days — aakashgupta · 2026-10-29
- Thiel Fellowship's wild ROI: 40 under-20 fellows founded 13 unicorns, 4 are billionaires — FinanceYF5 · 2026-10-29
- KUKA CEO on physical AI: how 600,000 deployed robots learn new tricks — lukas_m_ziegler · 2026-10-29