Anthropic threat report's untold side: Kimi/DeepSeek users' prompts silently routed to Claude
SiteSpecialist6295 · reddit · 2026-09-12
Reading Anthropic's new threat intelligence report, the author argues discussion fixates on "Chinese labs distilling Claude" while ignoring the bigger privacy story.
- Per Anthropic, Moonshot/Kimi and DeepSeek silently routed some users' requests to Claude without their knowledge, allegedly including internal company docs, live credentials, government-related data, and surveillance information.
- Conversely, Anthropic could identify campaigns, link huge account clusters, estimate sponsoring organizations, and analyze millions of exchanges—showing how much visibility an AI provider can have over traffic reaching its models.
- Conclusion: no winners. Users don't know where prompts end up; providers must build ever-stronger monitoring; routers obscure data flows. The real lesson may be that we're building an ecosystem where everyone watches everyone, while users have the least visibility of anyone.
More from AGI Musings
- Coding matters less than problem-finding: the shift from code engineers to product engineers — CtrlAltDwayne · 2026-10-28
- Rachel Thomas points to classic paper on articulating the value of not building tech — math_rachel · 2026-10-28
- Hypothetical $2T Anthropic IPO framed as channeling $100B+ a year into the real economy — sanjaykalra · 2026-10-28
- Biotech CEO: drug discovery will lag in the AI revolution, not lead it — vaibhavbetter · 2026-10-28
- AI cyber-wreckage noise far outstrips actual damage so far, argues new analysis — timigod · 2026-10-28
- 'Enjoy the low-polygon era': F-117 meme captures today's AI constraints — willcb · 2026-10-28