AI Summary
5 min readGoogle's Liz Reid, VP of Search, explains how AI is transforming the company's core product without undermining its link-based model or ad business. In this Odd Lots interview, she details mechanisms for integrating generative AI like AI Overviews and the Gemini app, user behaviors, query evolution, and future adaptations, emphasizing complementarity between AI and the open web.
AI's Long History and Current Integration in Search
Reid notes AI has underpinned search for years, from early spell correction to transformer-based models like BERT (2018) and MUM. Today, generative AI powers AI Overviews—summaries atop results—and AI Mode within search, alongside the separate Gemini app. Deployment isn't automatic; AI Overviews appear based on user signals indicating added value, such as for complex queries where a single page falls short. For "Corgi," users often seek images or breed details, favoring links over summaries; "What is a Corgi?" triggers an overview. Models improve coverage as they advance, prioritizing high-quality responses. This avoids "AI for AI's sake," mirroring decisions for features like weather boxes.
Caveat: Quality hinges on signals; poor model performance blocks overviews. Reid stresses surfacing expertise from individuals, as people still crave perspectives (e.g., fashion influencers) beyond quick facts.
Resolving the Click vs. Summary Tension
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
What you'll learn
- 1 (01:48) **AI Threat to Google Search** - Hosts introduce episode theme on AI disrupting Google's search and ad business
- 2 (06:46) **Liz Reid's Role at Google** - VP of Search explains leading product, engineering, design, and data science for core search
- 3 (07:16) **AI's Long History in Search** - Reid traces AI from spell correction to BERT, MUM, and now AI Overviews
- 4 (08:49) **AI Summaries vs Click-Out Business** - Addresses tension between instant answers and ad revenue from clicks
- 5 (11:54) **How AI Overviews Are Triggered** - Shown based on user signals of added value, not just questions
- 6 (16:55) **User Patterns Across Google Products** - Distinct query types: informational on search, creative/productivity on Gemini
- 7 (19:17) **Google as Fact-Checker for LLMs** - Users cross-check LLM outputs on Google, expanding to longer queries
+ Full timestamped outline available in the app
Show Notes
Not too long ago, search engines were the dominant form of querying the internet. But that's changing since the rise of large language models like ChatGPT, Claude, and Google's Gemini. More and more people are getting their online info through AI, effectively bypassing the search bars of old and creating a tension for large tech companies that offer AI models, but also make money from web traffic and search-related advertising. In this episode, we speak with Elizabeth Reid, VP of search at Google. Liz has been with the company for more than two decades, witnessing multiple tech transformations in that time. So we talk with her about how Google is incorporating Gemini into search via AI overviews, what that means for traffic and ad sales, and the practical experience of search in an age of LLMs and internet slop.
Subscribe to the Odd Lots Newsletter
Join the conversation: discord.gg/oddlots
See omnystudio.com/listener for privacy information.
More from this podcast
Odd Lots →