Sundar Pichai on AI, search’s future, and the web’s evolution
▼ Summary
– Sundar Pichai restructured Google after the ChatGPT moment, combining DeepMind and Brain into Google DeepMind and centralizing AI infrastructure to move faster.
– Pichai sees the future of Search as merging the intelligent search box with the Gemini Spark agent platform, enabling searches to trigger tasks rather than just return results.
– Pichai acknowledges that for subjective queries like “best Chromebook,” the AI overview can be too opinionated and sees room for improvement.
– Pichai addresses publisher concerns about declining search traffic, saying Google is committed to connecting users to high-quality content but notes the ecosystem is evolving.
– Pichai agrees with Demis Hassabis that we are in the “foothills of the singularity,” with AGI likely arriving sooner rather than later, though he avoids giving a specific timeline.
In a recent conversation following Google I/O, Alphabet and Google CEO Sundar Pichai sat down for his fifth annual post-conference discussion. The interview covered a wide range of topics, from internal restructuring to the future of search and the web. Pichai acknowledged that the arrival of ChatGPT prompted a significant rethinking of Google’s organizational structure, leading to key executive changes and a more aggressive posture.
When asked how Google is currently organized, Pichai described three main business pillars: Search, YouTube, and Google Cloud, supported by major platforms like Android and Chrome, all powered by the company’s AI and infrastructure teams. He emphasized that the AI era allows for a more unified approach, with a common infrastructure,centered on the Gemini model,enabling consistency across products. This, he argued, helps resolve the historical tension between Google’s culture of launching many products and the need for focus.
Pichai explained his decision-making framework, noting that most decisions are not truly consequential; what matters more is making them quickly to maintain organizational velocity. He contrasted this with a few high-stakes choices, like merging the Brain and DeepMind teams into Google DeepMind, which required careful deliberation. He also discussed how AI is already changing internal workflows, with engineers increasingly directing teams of AI agents rather than writing all code themselves.
The conversation shifted to the future of Google Search and the concept of “Google Zero”,the idea that Google traffic to websites could fall to zero as the company answers queries directly on results pages. Pichai pushed back on this notion, pointing to Google’s ongoing commitment to connecting users to web content. However, he acknowledged the evolving landscape, noting that publishers like Condé Nast are now publicly planning for a world with zero search traffic. Pichai responded by saying that while the information ecosystem is broader than just Google, the company remains dedicated to reflecting high-quality content in its products.
On the topic of AI agents, Pichai described them as the next evolution of the web. He highlighted new features like the Intelligent Search box and the Gemini Spark agent platform, which can perform tasks like booking tickets. He predicted these capabilities will eventually converge into a single, seamless user experience. When asked about the potential for AI to destabilize Google’s role as a common source of truth, Pichai distinguished between objective queries (e.g., “What is the capital of the USA?”) and subjective ones (e.g., “Plan a trip to Montreal”), arguing that personalization is natural for the latter.
Addressing public anxiety about AI, Pichai acknowledged that the rapid pace of change is unsettling for many people. He rejected the idea that this is merely a marketing problem, instead calling it a “multilayered” issue involving economic uncertainty, energy concerns, and societal adaptation. He emphasized the need for industry responsibility, government regulation, and public involvement as AI continues to roll out.
Finally, Pichai discussed the timeline to AGI (artificial general intelligence), agreeing with Google DeepMind CEO Demis Hassabis that we are “at the foothills of the singularity.” While he declined to give a specific timeline,three years, five years,he stressed that the rate of progress is so rapid that the exact label matters less than preparing for increasingly powerful systems. He noted that the core technology, based on transformers, continues to evolve, and that AI is already enabling novel capabilities, from scientific discovery to creating new operating systems.
(Source: The Verge)
