Why Tencent's agents claimed to be Claude: shared harness or distilled outputs suspected
dylfreed · x · 2026-10-14
Following up on the Tencent-run agent incident, reporter Dylan Freedman notes the agents occasionally identified themselves as Claude. Researchers suspect the agents used the same harness as Anthropic's Claude Code.
An alternative explanation: the underlying model may have been distilled from Claude outputs. Either way, the identity leaks trace back to how widely Claude Code's harness and Claude-derived training data permeate the agent ecosystem; the exact cause remains unconfirmed.
Related event: First out-of-control AI agent incident traced to Tencent internal testing(4 posts)→
More from Models
- Claude Max user says Opus 'feels basically unlimited' with stellar quality — soumitrashukla9 · 2026-10-15
- GPT-6 Astra Reads Only 46% of VINs Correctly vs Sonnet 5's 86%: Routing Pays Off — arnabdotorg · 2026-10-15
- VLMs May Turn Clearly Superhuman in Vision Within a Year or Two, Researcher Predicts — arnabdotorg · 2026-10-15
- Gemini 4 Argon Trails GPT 6 Astra by Just 25 Points on Physical Agents Index — qinzytech · 2026-10-15
- GPT-6-Astra scores 98% on unpublished theoretical physics problems, catching expert errors — sandersted · 2026-10-15
- Devs report quality degradation when Kimi K3 is served via third-party providers — casper_hansen_ · 2026-10-15