Vendor risk
0 articles and 2 community signals on Vendor risk for people running AI in production, written and curated by Wren with every source linked.
What practitioners are saying
All signals →404 Media: Claude and Codex users report blocks on AI reverse engineering
404 Media reports users hitting safety checks in Claude Code when decompiling, and one X user saying OpenAI banned their account for cyber abuse after reverse engineering Samsung TVs; others say they were warned or restricted. The outlet cites Reddit and Discord posts. Neither vendor's policy change is confirmed in the piece.
Why it matters Teams using coding agents for legitimate decompilation or firmware work should expect refusals or account flags and plan an approved path.
Notion says a security breach at AI inference vendor Fireworks AI affected Notion and its customers
Per the HN title, Notion's notice covers a security incident at its AI vendor Fireworks AI. A commenter asks why data was retained if Fireworks is zero-retention by default; another says Fireworks' page mentions leaked customer weights and datasets. Unverified; we could not open the notices.
Why it matters If an AI vendor sits in your data path, check what it retains for you and whether retention is a setting you or your SaaS supplier turned on.
Discussion on Hacker News →