
Google Part III: The AI Company. Google is amazingly well-positioned... will they win in AI? (Audio)
Hosted by Ben Gilbert and David Rosenthal
- 4h 3m
- Transcript
Summary
Written by a model from the transcript; not yet reviewed.
The episode traces Google's AI journey from a 2001 lunch conversation between Georges Herrick and Noam Shazer about compressing data as understanding, through the 2011 CAT paper that proved large neural networks could learn from unlabeled data, to the 2017 Transformer paper that became the foundation of the entire AI revolution. The hosts argue Google is in the most classic textbook case of the innovator's dilemma: they created the technology that spawned OpenAI, Anthropic, and Nvidia's trillion-dollar rise, yet all eight Transformer authors left to build competitors. The episode covers the DNN Research auction (Baidu's $12M bid, final price $44M), DeepMind's $550M acquisition, the TPU chip built in 15 months, and Google's response to ChatGPT with Code Red, Bard, and Gemini.
The second half examines Google's current position: $370B revenue, $140B earnings, a $3T market cap, 2-3 million TPUs, and Google Cloud at $50B revenue growing 30% annually. The hosts present bull and bear cases, noting Google is the only company with a frontier model, AI chip, hyperscaler cloud, and self-sustaining funding. Waymo is highlighted as a potential Google-sized business with 100M+ autonomous miles and 91% fewer serious crashes. The episode concludes that Google is threading the needle between protecting its search franchise and competing in AI, with the central question being whether they'll risk their profits to win.
- Google is the only company that has both a foundational frontier model and an AI chip at scale, making it structurally unique in the AI market. 3:40
- All eight authors of the 2017 Transformer paper left Google to start or join AI startups, including OpenAI, within a couple of years of publication. 2:00:48
- Waymo has driven over 100 million miles with no human driver and shows 91% fewer crashes with serious injuries compared to the average human driver. 3:00:12
- Google Cloud reached over $50 billion in annual revenue run rate, growing 30% year over year, making it the fastest-growing major cloud provider. 3:24:52
- Google's token processing went from 10 trillion in April 2024 to nearly 500 trillion in April 2025, a 50x increase in one year. 3:45:40
Highlights
- 41:10
CAT Paper Proves Unsupervised Learninginsight
Google Brain trained a nine-layer neural network on 16,000 CPU cores to recognize cats from unlabeled YouTube frames. A single neuron activated for cats without ever being told what a cat was, proving large networks could learn meaningful patterns without supervision.
- 52:07
DNN Research Auction at Harrah'sstory
Jeff Hinton ran a live auction from his hotel room at the NIPS conference in Lake Tahoe. Baidu opened at $12M, DeepMind dropped out for lack of cash, and the researchers stopped bidding at $44M, choosing Google.
- 1:21:52
Hinton's Custom Jet Harnessstory
Alan Eustace chartered a private jet and built a custom bed-and-straps rig so Jeff Hinton, who couldn't sit due to a back condition, could lie flat during takeoff and landing for the DeepMind diligence trip to London.
- 2:00:48
Transformer Paper Published Quietlyinsight
Google allowed eight researchers to publish 'Attention Is All You Need' in 2017. The paper has since been cited over 173,000 times, making it the seventh most cited paper of the 21st century, and all eight authors eventually left Google.
- 2:25:39
ChatGPT's Record-Breaking Launchdata
ChatGPT hit 1 million users within a week of its November 30, 2022 launch, 30 million by December 31, and 100 million registered users by January 23, 2023, making it the fastest product in history to reach that milestone.
- 2:35:34
Bard Launch and Stock Dropstory
Google rushed Bard to public availability in February 2023. The launch video showed an inaccurate factual response, stock dropped 8% that day, and the product was clearly inferior to ChatGPT, missing RLHF entirely.
- 3:00:12
Waymo's 91% Crash Reductiondata
A Waymo study released the month before the episode showed 91% fewer crashes with serious injuries or worse compared to average human drivers, even controlled for city surface streets only. The hosts note this addresses a cause of 40,000+ US deaths per year.
- 3:45:40
50x Token Processing Growthdata
Google went from processing 10 trillion tokens across all services in April 2024 to nearly 500 trillion in April 2025, and then to roughly one quadrillion by June 2025, illustrating the scale of their inference infrastructure.
About this episode
Google faces the greatest innovator's dilemma in history. They invented the Transformer — the breakthrough technology powering every modern AI system from ChatGPT to Claude (and, of course, Gemini). They employed nearly all the top AI talent: Ilya Sutskever, Geoff Hinton, Demis Hassabis, Dario Amodei — more or less everyone who leads modern AI worked at Google circa 2014. They built the best dedicated AI infrastructure (TPUs!) and deployed AI at massive scale years before anyone else. And yet... the launch of ChatGPT in November 2022 caught them completely flat-footed. How on earth did the greatest business in history wind up playing catch-up to a nonprofit-turned-startup? Today we tell the complete story of Google's 20+ year AI journey: from their first tiny language model in 2001 through the creation Google Brain, the birth of the transformer, the talent exodus to OpenAI (sparked by Elon Musk's fury over Google’s DeepMind acquisition), and their current all-hands-on-deck response with Gemini. And oh yeah — a little business called Waymo that went from crazy moonshot idea to doing more rides than Lyft in San Francisco, potentially building another Google-sized business within Google. This is the story of how the world's greatest business faces its greatest test: can they disrupt themselves without losing their $140B annual profit-generating machine in Search? *Links:* - Geoff Hinton’s 2007 Tech Talk at Google: https://youtu.be/AyzOUbkUf3M?si=sLQJsEwq-JEOQmIN - Our recent ACQ2 episode with Tobi Lutke: https://www.acquired.fm/episodes/how-to-live-in-everyone-elses-future-with-shopify-ceo-tobi-lutke - Worldly Partners’ Multi-Decade Alphabet Study: https://worldlypartners.com/businesshistory - In the Plex: https://www.amazon.com/Plex-Google-Thinks-Works-Shapes/dp/1416596585 - Supremecy: https://www.amazon.com/Supremacy-ChatGPT-Race-Change-World/dp/1250337747 - Genius Makers: https://www.amazon.com/Genius-Makers-Mavericks-Brought-Facebook/dp/1524742678 - All episode sources: https://www.acquired.fm/episodes/google-the-ai-company#sources *Carve Outs:* - F1: The Movie: https://www.imdb.com/title/tt16311594/ - Travelpro suitcases: https://travelpro.com - Glue Guys Podcast: https://open.spotify.com/show/18n6FcVMR5J0hSeboeETPc?si=b876358f752a474d - Sea of Stars: https://seaofstarsgame.co - Stepchange Podcast: https://open.spotify.com/show/3RHJ129caYkxh2FliqtaBP?si=2bef3a854e084a39 *More Acquired:* - Get email updates https://www.acquired.fm/email and vote on future episodes! - Join the Slack http://acquired.fm/slack - Check out the latest swag in the ACQ Merch Store https://www.acquired.fm/store! 0:00 The Innovator's Dilemma: Google's AI Challenge 4:36 From PageRank to language models: Google's Early AI Foundations 17:21 Parallelizing Language: Jeff Dean's Google Translate Revolution 24:05 Sebastian Thrun and Geoff Hinton: Laying Google X Foundations 35:10 Unsupervised Learning: Google Brain's Cat Paper Breakthrough 47:05 From AlexNet to DeepMind: The GPU AI Revolution 1:00:02 The DeepMind Acquisition: AlphaGo and AI's New Frontier 1:40:22 Building TPUs: Google's Hardware Bet on AI 1:51:25 "Attention Is All You Need": The Transformer's Impact 2:04:01 Elon's Exit, Microsoft's Entry: OpenAI's New Path 2:15:43 The Chat GPT Shockwave: Google's Code Red Response 2:37:32 Unifying AI: Gemini and the Google DeepMind Merger 2:46:58 From DARPA Challenge to Robotaxis: Waymo's Long Road 3:03:44 Waymo's Billion-Dollar Potential and Google's Financial Power 3:20:26 Google Cloud's Strategic Role in the AI Ecosystem 3:32:31 Bull and Bear Cases on the Future of Google 3:53:10 Navigating the Innovator's Dilemma: Google's AI Strategy 3:54:34 NFL Summit Carveouts, and Outro _Note: Acquired hosts and guests may hold assets discussed in this episode. This podcast is not investment advice, and is intended for informational and entertainment purposes only. You should do your own research and make your own independent decisions when considering any financial transactions._
Transcript
41,211 words · 2 speakers
Speaker names marked are inferred from speaking time and the show’s known hosts — they are not voice-verified.
Topics
- Artificial Intelligence
- Google Strategy
- Transformer Architecture
- Deep Learning
- AI Chips
- Autonomous Vehicles
- Cloud Computing
- Disruptive Innovation
- Waymo
Page updated · transcript and summary are machine-generated