Severe Recency Bias in Claude Opus's New Memory Feature
nptacek · x · 2026-07-09
Users report that Claude Opus exhibits a severe recency bias after enabling its memory function. The model overly references hypothetical content it generated in recent conversations, using it as a factual basis to answer completely unrelated new questions.
Users point out that the model treats its own outputs as absolute truth and lacks a verification mechanism for remembered information. This serves as a reminder not to blindly trust opaque, black-box memory systems.
Related event: Claude Opus Memory Feature Shows Severe Recency Bias(2 posts)→
More from Models
- Google says Gemini 4 has entered its most ambitious pre-training run yet — himanshustwts · 2026-07-22
- China’s AI arms race is increasingly defined by chips, data centers, and open models — BenBajarin · 2026-07-22
- Sam Altman is headed to Washington to brief Congress on OpenAI’s GPT-6 line — inductionheads · 2026-07-22
- Benchmark chart pits GPT-5.6 Luna, Grok 4.5 and Gemini 3.6 Flash on price and scores — iruletheworldmo · 2026-07-22
- Claim says Kimi was distilled from Fable, sparking a model-attribution jab — cephaloform · 2026-07-22
- Gemini 3.6 Flash is now available in Antigravity and chat — MartianOnJupiter · 2026-07-22