Podcast Summary: Intelligent Machines Episode IM 831: "9 Seconds of Google - Holy Widgets"

Release Date: August 7, 2025

Host: Leo Laporte | Guests: Jeff Jarvis, Paris Martineau, Vlad Prelovac

1. Introduction to Kagi Search Engine

The episode launches with Leo Laporte introducing Vlad Prelovac, the founder and CEO of Kagi—a search engine positioned as a superior alternative to Google. Vlad's journey to creating Kagi began in 2018 when he sought a reliable, ad-free search experience that respected user privacy.

[04:32] Vlad Prelovac: "We wanted to create a search engine that you can trust, aligning incentives to a paid business model where people pay for search results. If they're not satisfied, they simply leave."

2. Kagi's Business Model and User-Centric Approach

Kagi differentiates itself from Google by adopting a subscription-based model, eliminating ads and focusing solely on delivering high-quality, reliable search results. This approach fosters a positive feedback loop, encouraging continuous improvement based on user satisfaction.

[06:34] Vlad Prelovac: "Our users pay for high-quality search because they value it. This ensures we prioritize their interests over random advertisers."

3. Transition to a Public Benefit Corporation (PBC)

In 2024, Kagi transitioned to a Public Benefit Corporation, allowing it to balance profit motives with a mission to create a more human-centric and sustainable web.

[07:30] Leo Laporte: "Kagi is committed to creating a more human-centric and sustainable web that benefits individuals, communities, and society as a whole."

[07:49] Vlad Prelovac: "As a PBC, we aren't solely beholden to shareholders but also to our public benefit mission, influencing our strategic decisions and initiatives."

4. Advanced AI Integration Without Compromising User Experience

Kagi integrates Generative AI selectively, ensuring that AI enhancements are user-driven and not intrusive. Features include:

• AI Summaries: Accessible by appending a question mark to queries, providing concise overviews with proper citations.

  [11:53] Leo Laporte: "You don't get AI summaries unless you put a question mark. This keeps AI assistance optional and user-controlled."

• Kagi Assistant: A tool that leverages top Large Language Models (LLMs), grounded in Kagi's search results for enhanced accuracy.

  [12:55] Vlad Prelovac: "With Kagi Assistant, you access multiple LLMs integrated with our search, allowing for customized and accurate interactions."

5. Web Indexing and Enhanced Search Quality

Kagi employs a hybrid indexing strategy:

• Aggregated Search Results: Sourced from other search engines through paid integrations.
• Custom Index: Focused on non-commercial, unmonetized content such as personal blogs and forums, filling gaps left by mainstream search engines.

Additionally, Kagi employs algorithms to downrank pages overloaded with ads and trackers, promoting higher-quality content.

[15:35] Paris Martineau: "Can you talk a little bit about your web index and your news index?"

[15:39] Vlad Prelovac: "We focus on non-commercial content—personal blogs, forums—that offer genuine, unmonetized human writing, something often missing in other search engines."

6. User Privacy and Data Protection

Kagi prioritizes user privacy by minimizing data collection. It introduces Privacy Pass, a technology ensuring complete user anonymity even with paid accounts. This allows users to authenticate their subscriptions without revealing personal identities.

[27:47] Vlad Prelovac: "With Privacy Pass, we authenticate payments without knowing the user's identity, ensuring maximum privacy."

Additionally, Kagi offers a Tor address, enabling completely anonymous access.

[29:24] Leo Laporte: "For those deeply concerned about privacy, Kagi can be accessed via Tor, ensuring complete anonymity."

7. Challenges and Controversies: Cloudflare vs. Perplexity

A significant portion of the episode delves into the controversy between Cloudflare and Perplexity. Cloudflare accused Perplexity of using undeclared crawlers to bypass "no crawl" directives, violating web norms. Perplexity rebutted, asserting that their approach aligns with user-driven queries and does not involve traditional web crawling or data storage.

[25:34] Jeff Jarvis: "Perplexity fired back, stating they don't store information or train on it, unlike traditional crawlers."

[30:31] Leo Laporte: "Cloudflare set up dummy sites with strong crawl blocks and accused Perplexity of evading them, which Perplexity denies."

The discussion highlights the complexities of AI-driven web interactions and the balance between open web principles and respecting site-specific directives.

8. Expansion Plans and Future Developments

Vlad outlines Kagi's ambitious plans, including:

• Search API Launch: Expected within a month, catering to the rising demand in the AI age.

  [29:42] Vlad Prelovac: "Our Search API will significantly enhance AI model accuracy by providing high-quality search results."

• Browser Development: Expansion of their Orion browser to Linux, following successful releases on Mac and iOS.

  [30:48] Vlad Prelovac: "We're now working on Orion for Linux, expected in about nine months."

9. Broader Reflections on the State of the Web

Vlad expresses concerns over the "inshittification" of Google's search results due to ad-driven incentives, leading to a proliferation of low-quality, ad-centric websites. He emphasizes Kagi's role in restoring quality and trust in web search.

[18:41] Vlad Prelovac: "The last 20 years have deteriorated the web by prioritizing ads over content quality. Kagi aims to reverse this trend."

10. Closing Remarks and Community Engagement

Leo, Jeff, and Paris commend Vlad and Kagi's efforts, encouraging listeners to try the service and support its mission. They highlight the importance of user agency in shaping the future of search and AI technologies.

[34:31] Jeff Jarvis: "53,000 members and counting. Try Kagi for free and support a search engine that prioritizes your interests."

Notable Quotes:

• Vlad Prelovac [04:32]: "We wanted to create a search engine that you can trust, aligning incentives to a paid business model where people pay for search results. If they're not satisfied, they simply leave."

• Leo Laporte [06:34]: "Kagi is committed to creating a more human-centric and sustainable web that benefits individuals, communities, and society as a whole."

• Vlad Prelovac [11:55]: "AI has been part of search for decades... we're treating AI as a tool, on demand only, not pushing it down users' throats."

• Vlad Prelovac [14:51]: "Kagi is all about user agency. Putting the user in control, not figuring out for them what they want to see."

• Vlad Prelovac [18:41]: "The last 20 years have been built on a lie and it's now unraveling with these AI and using Google results."

Conclusion:

Episode IM 831 of Intelligent Machines offers an in-depth exploration of Kagi Search Engine, highlighting its commitment to user privacy, quality search results, and ethical AI integration. Through Vlad Prelovac's insights, listeners gain a nuanced understanding of the challenges posed by ad-driven search models and the potential for subscription-based, user-centric alternatives. The episode also touches on broader AI controversies and the evolving landscape of web technologies, emphasizing the critical role of user agency in shaping the intelligent future.