Hacker News Reader

The big question is still - how will the models reach production?

You can check out my blog post on publishing models to Docker Hub using Docker Model Runner :) https://www.docker.com/blog/publish-ai-models-on-docker-hub/

Well, more wondering how I go from having local models into a production app with the model available - I guess swap it with a cloud offering?

You can run the same model in production, using Docker Model Runner in Docker CE. This is assuming, the model works for your production use case :)