This message was deleted.

We have a blog post here that describes our approach at a high level: https://www.pulumi.com/blog/leveling-up-pulumi-ai-with-the-pulumi-registry/ I don't believe the Proxmoxve provider is currently in its vector database; so you're seeing the result of the model trying to answer "as best as it can". We plan to add our community providers, but we're being careful to make sure that adding more data does not cause a regression with our most popular queries.