Why Did Yahoo Build Its Own Search Engine Again After 17 Years?

Jim Lanzone says Yahoo Scout is not a frontier model like Gemini, ChatGPT, or Claude, and it is not meta-search in the way early Perplexity was, sitting on other engines the way Dogpile once did. Scout is a product on Yahoo's own data. Yahoo is still the third-ranked search engine in the United States, with 30 years of search history, publisher relationships, a knowledge graph, and what it knows about a user. Scout grounds that data, sends it to a lightweight model, and Yahoo renders the result. At launch the model was Anthropic's Haiku. Yahoo uses others now. The answers are unique because the data is Yahoo's.

The strategic reason was sharper. For 17 years Yahoo outsourced search, largely to Microsoft, who is still a partner. That left one of the most important products and revenue streams in someone else's hands. Owning it again was, in his words, life-changing. He spent his first ten years in search and cares about results that go past ten blue links. Yahoo also stopped competing with its publishers and went back to sending them traffic, which he treats as the job.