Users Complain Claude IQ Drop, Anthropic Scrambles to Recover
annetgriffin · x · 2026-08-19
Multiple users recently reported severe degradation in Claude model performance, specifically noting Fable 5 is "borderline unusable." This follows three major bugs in Claude Code during Aug/Sept 2025 that caused user churn. While issues were fixed before the 2025 holidays, sparking a new wave of building, Anthropic faces intense competition. They must now figure out how to retain model strengths while fixing issues driving users away, relying partly on earlier enterprise deals as a buffer.
More from Companies & People
- SAIL Lab Critiques Economist's Article on Acemoglu as Unconvincing — ChenhaoTan · 2026-08-19
- User claims OpenAI referral credits vanished after confirmation, support criticized — Hvung · 2026-08-19
- Rillet hits $1B valuation with $100M Series C as CFOs ditch SAP for AI agents — mattturck · 2026-08-19
- Inside Unitree's long road to IPO: Sequoia's 8-point memo and a robot dog on a sleeper train — elsewhere别处发生 · 2026-08-19
- Suits Are Back: The Boomer-Maxxing Strategy in AI — Linahuaa · 2026-08-19
- AI Sandbox Escapes Are Not Selling Points for Enterprise CIOs and Security — Miles_Brundage · 2026-08-19