To take your example: If I ask an LLM model like Claude or OpenAI who makes the best skates, their chatbot can add a link to Acme in its response. Said chatbots will be available for "free" just like Google search (I use free ChatGPT for instance).

If people are going to start asking LLMs commercially relevant questions IN LIEU of traditional Google search, the advertisers will start moving their money to the chatbot providers. Gemini will be one of those providers but only one among many. ChatGPT as verb takes on greater significance in this world, of course.

If I use the LLM but then go to Google for my purchase-driven search, Google is in much better shape. We're basically talking about two kinds of searching, just like one might search Instagram or Youtube for one class of things(e.g., how to skate or how to apply makeup) but Google search for actually purchasing your next pair of skates.

So a critical question is whether and/or how competitors like OpenAI pull users away from Google Search. That is a question yet to be decided. One interesting study I came across by way of MBI Deep Dives tried to test the issue as follows:

semrush.com - Google usage after chatgpt adoption

Based on the data, ChatGPT adoption did not reduce Google Search usage.

This supports the expansion hypotheses, meaning people did not substitute their typical Googling with ChatGPT-ing.

In fact, there was even a slight increase in average Google Search usage after ChatGPT adoption.

Of course, the above also seems consistent with Google's last couple earnings reports, which have yet to show a real weakness in its search business.