gpt.icon - Advances in AI technology could significantly lower costs, making the integration of Google search and AI realistic, and making search as smart as AI. - The threat to Google Search is that the declining cost of AI will make the subscription model much cheaper and ultimately make free services feasible. - The biggest risk for Google is that customers will flow to other companies before implementing AI integration, and successful AI integration will reduce the risk of customer exodus and maintain market dominance.

kazuho Maybe Google is thinking the other way around, and when the cost is 1/10th of the cost, it becomes realistic to provide ad-based services, and Google search and AI will be integrated and as smart as AI. Google search and queries to AI will be integrated and become as smart as AI. _sushiy: 10x faster means that Ai’s cost could be 1/10th and ChatGPT subscriptions could be around $2. Once it’s free, Google search and the like will really die, it’s a horrible time.

kazuho This story came out last year.

kazuho: I see, it’s an asset from a while ago, but if we were to add GPT-3 response to Google search, the cost would increase by 30%, enough to blow all the gross profit (but not enough to make a loss). On the other hand, for bing, even with the cost increase, if the search market share increases, the loss would decrease. The Inference Cost Of Search Disruption – Large Language Model Cost Analysis

_sushiy You can already make it so that it also produces query results to Gemini while searching, but I personally don’t use it at all, partly because Gemini is not very smart. I personally don’t use it at all. Even if they eventually catch up to ChatGPT, if other companies are able to return the same level of results as Google, Google’s overwhelming advantage will be destroyed.

kazuho I think the biggest risk for Google was that customers would flow to other companies before they were able to integrate AI. I think that if they integrate and it becomes a battle of AI vs. AI, the possibility of customer exodus will be reduced, and they will also have an advantage in the amount of computation (which will affect the quality difference of AI) and R&D competition based on the difference in market share. (I believe Google’s dominance is not in search quality, but in the oligopoly of the search and advertising business itself.)


This page is auto-translated from /nishio/AI技術の進化と検索市場への影響 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.