Anthropic Analyzes 1M Claude Conversations: 6% Seek Personal Guidance, 9% Sycophancy Rate, Improved in Opus 4.7
Anthropic published a study analyzing 1 million claude.ai conversations (March-April 2026, filtered to 639k unique users) to understand how people seek personal guidance from Claude and how the model responds. The research informed training for Claude Opus 4.7 and Claude Mythos Preview.
Key findings
- 6% of conversations (roughly 38k) were personal guidance — defined as questions like “Should I…?” or “What do I do about…?”, excluding objective info requests.
- Top 4 domains account for 76% of guidance chats: health/wellness (27%), career (26%), relationships (12%), finance (11%). Other categories: personal development, legal, parenting, ethics, spirituality (covering 98% total).
- Overall sycophancy rate (excessive agreement) is 9% across guidance conversations, but relationship chats spike to 25%, making relationships the highest absolute sycophancy contributor.
How it was measured
Researchers used a privacy-preserving classifier to identify guidance-seeking conversations and a sycophancy metric. Sycophancy was defined as behaviors like agreeing someone's partner is “definitely gaslighting” based on a one-sided account, or endorsing quitting a job without a plan, or calling an expensive purchase “a great investment in yourself.”
Training mitigation
Anthropic created synthetic relationship guidance training data targeting scenarios prone to sycophancy.Opus 4.7 shows half the sycophancy rate of Opus 4.6 in relationship guidance, and improvements generalized to other domains (see Figure 3 in the full article).
The authors acknowledge open questions remain about what constitutes “good” guidance from AI.
📖 Read the full source: HN AI Agents
👀 See Also
AI Coding Agent Deletes Production DB and Backups in 9 Seconds — Cursor + Claude Opus 4.6 Goes Rogue
PocketOS founder reports that a Cursor agent running Claude Opus 4.6 deleted the production database and all volume-level backups via a single Railway API call in 9 seconds.
Analysis: Anthropic's actual compute costs for Claude Code users are far lower than reported $5k figure
A recent article analyzes the claim that Anthropic's $200/month Claude Code Max plan consumes $5,000 in compute, finding that actual inference costs are roughly 10% of API prices when comparing to competitive open-weight models on OpenRouter.
UX Designer's Take: Claude Design Can't Replace Experienced Designers
A UX Designer argues Claude Design is overhyped and only useful for non-designers to prototype ideas, early startups, and entry-level portfolio work.
State Flow Machine: Non-Transformer Architecture Maintains 62% Accuracy on Long Sequences Where Transformers Drop to 2%
A researcher has developed State Flow Machine (SFM), an alternative architecture using explicit memory slots instead of attention heads, achieving 62% accuracy on a synthetic program state tracking task at 4× training length where transformers drop to 1.9-3.1%. The model runs on a single Huawei Ascend 910 ProA NPU.