cluster 'title' generated by ai

[danielroe]

we'll also need to add a lengthier cache for each cluster api response as this will become costly

[rijkvanzanten]

This would be very useful to have. I suppose it'd need some lightweight llm to summarize the various recognized issue content right? If so, would you lean towards integrating with a hosted model (w/ potential privacy / cost concerns) or try to include a embedded self-hosted model a la ollama (which might be huge)?

[danielroe]

I'm not worried about the privacy content as it's indexing publicly available github issues (and that's how we're getting the embeddings anyway). possible models include https://developers.cloudflare.com/workers-ai/models.