Google's Chief Technologist on Intelligent Search in the Age of AI
June 30, 2026
Search began as a way to find pages. AI is turning it into a way to ask, reason, decide, and act.
Search has always been more than a technical problem. It is a way of organising knowledge, connecting intent with information, and increasingly, turning questions into actions. In the age of artificial intelligence, that basic function is being redefined.
In this episode of TechSurge, host Sriram Vishwanath speaks with Prabhakar Raghavan, Chief Technologist at Google, about the long arc of search: from the early web and link analysis to knowledge graphs, language models, transformers, Gemini, and the unresolved question of how AI will change the way we find, trust, and use information.
Prabhakar reflects on his career as a computer scientist, researcher, and technology leader, beginning with his time at IBM Research, where he worked on algorithms, optimization, databases, and early information retrieval. He explains how the explosion of unstructured data on the web created a new class of technical and economic problems. Search was not simply about indexing pages; it was about imposing structure on a chaotic information environment and building mechanisms that could connect supply, demand, relevance, authority, and trust.
The conversation traces how early search evolved through link analysis and PageRank, drawing on ideas from scholarly citation analysis, graph theory, and algorithmic ranking. Prabhakar describes why authority and trust became central to search as the web grew, and why users themselves changed alongside the technology. As search engines became more capable, people moved from looking for simple webpages to asking richer, more contextual questions that required intent understanding rather than mere document retrieval.
Sriram and Prabhakar then explore the transition from classical search to AI-infused products. Through examples such as Gmail Smart Reply, Smart Compose, Google Drive recommendations, and knowledge graphs, Prabhakar shows how prediction, context, and language modelling were already reshaping user experiences well before the current generative AI wave. These systems were early signals of a broader shift: computers moving from retrieving information to anticipating what users might need next.
The episode also offers a technical tour of the major algorithmic milestones that led to today’s AI systems, including deep learning, sequence-to-sequence models, attention mechanisms, transformers, and the compute architectures needed to train and serve large models. Prabhakar explains why attention changed the quality of language modelling, why AI systems appear increasingly conversational, and why compute remains one of the central constraints in the field.
At the heart of the discussion is the central tension facing search today: if AI systems can generate answers directly, what becomes of search as we know it? Prabhakar does not frame AI as the end of search, but as its next transformation. The future of search may be less about finding a page and more about understanding intent, synthesising knowledge, reasoning through ambiguity, and helping users complete complex tasks.
The conversation closes with deeper questions about AI world models, hallucination, test-time compute, diffusion models, recursive self-improvement, theorem proving, and whether AI systems can ever reason with the same grounded understanding as humans. For Prabhakar, the challenge is not only to build more powerful models, but to understand their limits, failure modes, and relationship to truth.
This episode is a wide-ranging exploration of how search became one of the defining technologies of the internet age—and how artificial intelligence may now force us to rethink what it means to search at all.
Sign up for our newsletter at techsurgepodcast.com for updates on upcoming TechSurge Live Summits and future episodes.
Links:
- Prabhakar Raghavan - Google Research profile: https://research.google/people/prabhakarraghavan/?&type=google
- Prabhakar Raghavan - Google blogs and writing: https://blog.google/authors/prabhakar-raghavan/
References Mentioned During the Discussion
- Brin and Page - The Anatomy of a Large-Scale Hypertextual Web Search Engine: https://research.google/pubs/the-anatomy-of-a-large-scale-hypertextual-web-search-engine/
- Page, Brin, Motwani and Winograd - The PageRank Citation Ranking: https://ilpubs.stanford.edu:8090/422/1/1999-66.pdf
- Jon Kleinberg - Authoritative Sources in a Hyperlinked Environment: https://www.cs.cornell.edu/info/people/kleinber/auth.pdf
- Manning, Raghavan and Schutze - Introduction to Information Retrieval: https://nlp.stanford.edu/IR-book/
- Google - Introducing the Knowledge Graph: things, not strings: https://blog.google/products-and-platforms/products/search/introducing-knowledge-graph-things-not/
- Google Help - How Google's Knowledge Graph works: https://support.google.com/knowledgepanel/answer/9787176
Further Reading
- Krizhevsky, Sutskever and Hinton - ImageNet Classification with Deep Convolutional Neural Networks: https://proceedings.neurips.cc/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html
- Vaswani et al. - Attention Is All You Need: https://papers.neurips.cc/paper/7181-attention-is-all-you-need
- Devlin et al. - BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding: https://aclanthology.org/N19-1423/
- Chen et al. - Gmail Smart Compose: Real-Time Assisted Writing: https://arxiv.org/abs/1906.00080
- Kannan et al. - Smart Reply: Automated Response Suggestion for Email: https://arxiv.org/abs/1606.04870
- Hoffmann et al. - Training Compute-Optimal Large Language Models: https://arxiv.org/abs/2203.15556
- Tay et al. - Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers: https://research.google/pubs/scale-efficiently-insights-from-pre-training-and-fine-tuning-transformers/
- Wei et al. - Chain-of-Thought Prompting Elicits Reasoning in Large Language Models:https://proceedings.neurips.cc/paper/2022/hash/9d5609613524ecf4f15af0f7b31abca4-Abstract-Conference.html
- Vafa et al. - Evaluating the World Model Implicit in a Generative Model: https://proceedings.neurips.cc/paper_files/paper/2024/hash/2f6a6317bada76b26a4f61bb70a7db59-Abstract-Conference.html
- Google DeepMind - AlphaGeometry: An Olympiad-level AI system for geometry: https://deepmind.google/blog/alphageometry-an-olympiad-level-ai-system-for-geometry/
- Trinh et al. - Solving Olympiad Geometry Without Human Demonstrations: https://www.nature.com/articles/s41586-023-06747-5
- Google DeepMind - AI solves IMO problems at silver-medal level: https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/
- Li et al. - A Survey on Deep Learning for Theorem Proving: https://arxiv.org/abs/2404.09939
- Grossman et al. - How Generative AI Disrupts Search: https://arxiv.org/abs/2604.27790
- Xu, Iqbal and Montgomery - Measuring Google AI Overviews: https://arxiv.org/abs/2605.14021
- Aral, Li and Zuo - The Rise of AI Search: https://arxiv.org/abs/2602.13415
Timestamps:
[00:00:00] Highlights
[00:01:23] Podcast Sponsor
[00:02:50] Introduction
[00:04:08] From Researcher to Operator
[00:06:40] Why Search Became the Core of the Web Economy
[00:10:52] The Explosion of Unstructured Data
[00:12:13] The Market Mechanism Behind Web Search
[00:14:31] Could Anyone Have Predicted Where Search Was Going?
[00:15:24] Yahoo, Google, and the Technical Problem of Search
[00:16:16] PageRank, Link Analysis, and Trust on the Web
[00:18:24] From Citation Analysis to User Behaviour
[00:20:49] Search Was Never Just About the Query
[00:24:58] Larry, Sergey, and the Early Days of Google
[00:26:20] Gmail, Smart Reply, and Early Generative AI
[00:25:10] From Smart Reply to Smart Compose
[00:26:25] Predicting What Users Need Next
[00:29:09] Why Google Saw Context Differently
[00:38:06] The Long Arc from PageRank to Transformers
[00:41:30] From Computer Vision to Machine Translation
[00:42:11] Why Attention Changed AI
[00:44:09] How the Attention Mechanism Works
[00:45:26] Why AI Became a Compute Problem
[00:47:42] Why Better AI Is Not Just Better Token Prediction
[00:49:23] What Comes After Transformers?
[00:50:16] Diffusion Models and the Cost Curve
[00:55:28] How AI Learning Differs from Human Learning
[00:56:37] What AI Still Doesn’t Understand
[00:59:22] Test-Time Compute and Reducing Hallucinations
[00:50:00] Will AI Remain Compute-Constrained?
[01:02:09] Recursive Self-Improvement and Theorem Proving
[01:04:02] Could AI Discover Relativity or Ramanujan’s Identities?
[01:05:07] Is Google an AI-Native Company?
[01:06:49] Physical AI, Waymo, and Robotics
[01:08:38] What Has AI Already Proved Us Wrong About?
[01:10:01] The One AI Prediction He Is Willing to Make
[01:12:01] Advice for Young Entrepreneurs in the AI Era
[01:14:56] Why Fundamental Science Still Matters
[01:15:56] Closing Remarks
.png)
The Race to Build the Next Trillion-Dollar AI Chip Company
Almost 2% of U.S. GDP will be spent on AI infrastructure this year, nearly double 2025's figure. But beneath those headline numbers, the composition of that spending has quietly flipped: for the first time, dollars spent on running models in production now outweigh dollars spent training them.
In this episode of TechSurge, host David Goldman speaks with Austin Lyons, a semiconductor analyst at Creative Strategies, co-host of the Semi Doped podcast, and author of the Chipstrat newsletter. Lyons previously worked as a hardware engineer at Intel and as a product manager on John Deere's autonomous tractor and Blue River Technology teams before turning to full-time chip industry analysis.
The conversation opens with why AI buyers have moved from assembling commoditized parts to buying entire pre-integrated systems, tracing how Nvidia's rack-scale approach, exemplified by its 72-GPU Grace Blackwell racks, made turnkey deployment the default, and why that raises the bar for any chip startup trying to compete. Lyons and Goldman then unpack how inference workloads have split into two distinct problems, prefill and decode, and how that split created an opening for SRAM-based challengers to outperform general-purpose GPUs on decode speed.
From there, the discussion turns to the rise of neoclouds, the GPU-rental companies that grew into public businesses worth well over $100 billion combined, and why so many traditional investors missed them. Lyons and Goldman work through the circular financing debate head-on: the mechanics of Nvidia's equity stakes, GPU-backed debt, and hyperscaler off-take agreements that critics compare to dot-com-era vendor financing, and the counterargument that demand is simply outrunning fixed supply.
The episode closes on Lyons's own framework for identifying the next trillion-dollar chip company, built on four conditions including the ability to run trillion-parameter models at rack scale, beat an incumbent on a key performance metric, and land a frontier anchor customer, along with a look at how AI-assisted chip design is lowering the barrier for more companies, from OpenAI to electric vehicle makers, to design their own custom silicon.
Sign up for our newsletter at techsurgepodcast.com for updates on upcoming TechSurge Live Summits and future episodes.
.png)
The Nobel Winner Who Built Google's Quantum Computer: Why I'm Building NVIDIA of Quantum
In this episode, Nobel Prize-winning physicist Dr. John Martinis reveals how his breakthrough in superconducting qubits made quantum physics real at macroscopic scale and what it means for the future of technology. The former lead of Google's Quantum AI lab explains why quantum computing is so fragile, why a lot of hype has a low chance to work, and why his fabless company Qolab could be the Nvidia of quantum computing.
In this conversation, Dr. Martinis joins Tech Surge to explain the science behind macroscopic quantum coherence, the engineering challenges of scaling quantum computers, and how hybrid quantum-classical computing will shape the future of technology.
The conversation covers:
✅ How the superconducting qubit breakthrough won the Nobel Prize in Physics
✅ Why Nature wants to destroy quantum coherence and why quantum is fragile
✅ From academic physics to building Google's quantum computer
✅ The engineering challenge of scaling quantum computing beyond the lab
✅ Why a lot of quantum computing hype has a low chance to work
✅ How Qolab's fabless model could scale quantum hardware
Sign up for our newsletter at techsurgepodcast.com for updates on upcoming TechSurge Live Summits and future episodes.
Intel CEO Lip-Bu Tan on 40 Years of Contrarian Bets in Semiconductors
Silicon Valley was built on semiconductors, but for nearly two decades, venture capital shifted its attention towards software. Today, AI is changing that as the demand for compute, memory and networking explodes, hardware is once again at the centre of the industry's biggest bets.
In this episode of TechSurge, host Michael Marks speaks with Lip-Bu Tan, CEO of Intel and one of the semiconductor industry's most influential investors and executives. The conversation traces Tan's journey from studying nuclear engineering at MIT to leading Cadence's turnaround, investing in more than 500 technology companies, and now steering Intel through one of the most significant transformations in its history.
Tan shares his VC conviction on backing semiconductor startups when most venture investors favored software, and why he believes AI's next breakthroughs will come from advances in memory, packaging, photonics, cooling and high-speed connectivity. He also opens up on the leadership philosophy that defined his time at Cadence, where listening to customers and building a culture of responsiveness became the foundation of the company's revival.
Wearing his CEO hat, Tan explains Intel's long-term strategy, why vertical integration still matters, how the company plans to reconnect with the startup ecosystem, and why missing another technology wave is not an option.
Sign up for our newsletter at techsurgepodcast.com for updates on upcoming TechSurge Live Summits and future episodes.