It would do Google good to publish an AI integration philosophy to address critical issues related to AI in search. Such a document would allow for public scrutiny, which is crucial when your user base is a few billion strong and you are as influential player in the field. It would also serve as a good platform for other companies to study. Kagi has an AI integration philosophy [1], which, while modest, has successfully navigated us through the current AI cycle and I think helped our users understand what they can expect from us.
AI can enhance the search experience, but the core search functionality must be strong and robust first. An LLM grounded in search will only be as good as the search results it gets.
Google's "all-in" rollout of AI into search without first addressing these foundational principles and improving the core search experience is interesting - does it really mean they are betting the house on AI? The more I see things like this from Google and others, the more I want to go back to basics and just continue improving the core search quality.