Core Viewpoint - The article discusses the emerging demand for specialized search capabilities designed for AI, highlighting a fundamental shift in search from human-centric products to digital infrastructure that supports AI operations [1][4]. Group 1: Transition of Search - Search is undergoing a transformation to become a foundational infrastructure for AI, similar to how cloud computing supports the internet [1][4]. - AI products like Figma, Cursor, and Notion are evolving from static tools to interactive entities that can engage in dialogue [3][4]. - The integration of search into AI products is at varying stages, with companies like Cursor and Notion still in early development [4][9]. Group 2: New World Demands - The fragmentation of search will occur as each product develops its own specialized search needs, focusing on speed, quality, and the nature of results [6]. - Traditional search engines profit from clicks, while embedded search will generate revenue based on the quality of results provided [7]. - The quality of search retrieval will become a key differentiator, prioritizing recall rates and structured data over ad-filled results [8]. Group 3: Opportunities in Search - Opportunity 1: Providing real-time web search for large language models (LLMs) through optimized search engines like Exa, which focus on enhancing LLM performance [10][11]. - Opportunity 2: Enabling deep research capabilities for humans, surpassing traditional search engines, with tools like OpenAI and Exa's Websets [12]. - Opportunity 3: Offering private data search solutions for enterprises, unlocking knowledge trapped in SaaS platforms, exemplified by Glean's growth [13]. Group 4: Future Predictions - Search APIs are expected to thrive, with valuable search companies operating as invisible infrastructure for new AI applications [14]. - A fragmented search ecosystem will emerge, with numerous winners, while Google is likely to maintain its dominance in consumer search [15]. - The addressable market includes billions of knowledge workers, with pricing models shifting towards subscription-based systems that enhance productivity [16][17].

Core Viewpoint - The article discusses the fundamental shift in search technology as it transitions from a product used directly by humans to a digital infrastructure that supports AI operations, highlighting the emerging demand for AI-specific search capabilities [1][2]. Group 1: Transition from Product to Infrastructure - AI is increasingly integrated into software tools, transforming them from static applications to interactive products that users can converse with [1]. - As AI becomes embedded in every product, the need for specialized search capabilities arises, as AI cannot retain all information and requires tools to access vast and dynamic data [2]. Group 2: New World Demand - The shift to search as infrastructure will lead to fragmentation, with each product having its own search needs, resulting in specialized search tools tailored to various requirements [3]. - New revenue models will emerge, with embedded search generating income through results rather than clicks, shifting power to pure search service providers [4]. - The quality of search retrieval will become a key differentiator, focusing on recall rates and structured data rather than ad-filled results [5]. Group 3: Opportunities in AI Search - Opportunity 1: Providing real-time web search for large language models (LLMs) through APIs, with companies like Exa building AI-optimized search engines [6]. - Opportunity 2: Enabling deep research capabilities for humans, surpassing traditional search engines, as demonstrated by OpenAI's offerings [7]. - Opportunity 3: Facilitating private data searches for enterprises, addressing knowledge locked in SaaS platforms, with Glean showing significant momentum [8]. Group 4: Addressable Markets and Benefits - Addressable market includes every product utilizing large language models, with pricing based on usage rather than ad bidding, providing high-quality real-time knowledge [9]. - Addressable market includes millions of knowledge workers, with subscription pricing based on time savings, significantly reducing research time [10]. - Addressable market exceeds $400 billion in enterprise productivity, with pricing based on seat counts, enhancing employee productivity and preserving institutional memory [11]. Group 5: Future Predictions - Search APIs are expected to thrive, with valuable search companies potentially operating without visible result pages, serving as invisible infrastructure for new applications [12]. - A fragmented search ecosystem will emerge, with numerous winners, as API search infrastructure powers millions of products, while Google continues to lead in consumer search [12]. - A more informed world will develop as search becomes embedded in every application and context, enhancing the credibility and intelligence of products [13].