The BaseOfficialCloudflare
Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
A Cloudflare blog post, shared on Hacker News, is titled as describing cross-request KV cache exposure in the GLM-5.3 model on Workers AI, affecting 155 responses. We saw only the title; the cause, the data exposed and the fix are not known to us.
Why it matters
The same thing, for different readers.
What it means for you
Cloudflare's post title suggests that in 155 responses an AI service showed data from other requests, which matters for anyone sending private text to hosted models.
For lawmakers
The post title points to a disclosed leak between users of a hosted AI model, a case for rules on incident reporting by AI providers.
The record
- Questions
- Are AI companies putting stronger safeguards into practice?
- Also covered by
- Hacker News
- What we read
- We have read only the description the publisher gives. No quotations until someone reads the source itself.
- Found by
- The Radar, run of October 27, 2026, via
hacker-news-50-points(tier 3: discovery only) - Cite this
- Cloudflare, “Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)”, October 26, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/cloudflare-2026-10-26-glm-kv-cache-exposure
Our notesLight lane: nothing fetched and the excerpt was empty, so this rests on the title alone. What was exposed, how, and whether it is fixed are unknown. Read the post before publishing; change kind and lenses may need to change.