The community members are discussing Google's use of Retrieval Augmented Generation (RAG) in their generative AI search experience. One community member explains that it involves vector search and response synthesis, and provides a link to an example of how to build a similar pipeline using the llamaindex modules. Another community member wonders if it would be beneficial from an SEO perspective to align with Google's embedding model to optimize for the context of individual searches. The discussion also touches on the possibility of Google embedding the HTML of all websites, which one community member finds surprising, but another suggests is likely due to Google's consistent crawling of websites.
@jerryjliu0 thanks! Interesting I wonder then from a SEO perspective it make sense to align with their embedding model to optimize being the “most similar” for context an individual is searching for
