Is Search Dead?
Learn if LLMs are taking over search and what it means for SEO and AEO.
With the epic rise in popularity in ChatGPT and the use of LLMs it is increasing popular to shout from the rooftops that “Search is Dead!”. However the true story is more complicated than that as both traditional search and generative answer engines are developing faster and faster.
In the beginning, the battle between ChatGPT and Google could be summarized as retrieval versus generation. A Google search would provide you with a list of links based on its foundational algorithms such as PageRank, and ChatGPT 3 would provide you with a generated text answer based solely upon the training data used - a snapshot of the internet from many months prior.
The downsides of a Google search being a list of results which may or may not answer your query. The downsides of GPT 3 being that the responses based upon the training data could be out-of-date or unexpectedly biased. Thus an accepted truth was born. In order to discover accurate results, you must use a search engine. In order to discover answers quickly, you must use a Large Language Model.
Wait… LLMs search too, right?
In October 2024, this all changed. ChatGPT Search was introduced, allowing users to specify they wanted to have the generated answer refer to an external web search that the answer from ChatGPT could request.
However, Google had pre-empted these advancements by launching AI Overviews which at first received a lukewarm reception given several examples of comical suggestions to use glue on pizza, and to eat rocks - but only small ones!