Ask HN: Do you know what data your AI coding agent sends to the cloud?
The article discusses concerns about data privacy when using AI coding agents. Users express trust in the tools but remain cautious about sensitive information being sent to the cloud. There are policies in place at some workplaces to restrict sensitive data to on-premises models.
- ▪Users generally do not send sensitive information to AI coding agents.
- ▪Some workplaces have policies restricting sensitive data to on-premises models.
- ▪There is no evidence of cloud providers misusing chat history beyond storage.
Opening excerpt (first ~120 words) tap to expand
I trust the tool in that I don't send anything sensitive in there! Unless I built it, I assume it's going somewhere.We have a policy at work around this where our most sensitive data can only be passed to on prem models.That being said, I have no evidence of anything going to the cloud or frontier providers doing anything with chat history other than storing it for later.Self-hosted + custom harness for anything I don't want getting out at all.
Excerpt limited to ~120 words for fair-use compliance. The full article is at Ycombinator.