Oct 28, 2025 · 26m · big-technology

Google Research Head Yossi Matias: AI For Cancer Research, Quantum's Progress, Researchers' Future

Yossi Matias · 21m spoken Alex Kantrowitz · 2m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

In this episode of the Big Technology Podcast, host Alex Kantrowitz interviews Google Research Head Yossi Matias on the transformative roles of generative AI and quantum computing. Matias articulates how AI acts as an amplifier of human ingenuity in scientific discovery and explains Google's philosophy of connecting foundational research with real-world impact.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 12.2% of the talking time here. How this is scored →

Alex as informed peer 4.5 Guest teaching 4.2 Guest disagreement 1.2 Alex pushing back 3.3
05100:0010:0020:003:10–6:54 · Alex as informed peer 4/10 Quantum Computing Breakthroughs and Long-Term Real-World Impact Kantrowitz asks an informed question comparing recent headline breakthroughs with the recurring 5-to-10-year timeline disconnect. Matias explains Nature benchmark verifications and sets expectations for practical application within five years.6:55–10:40 · Alex as informed peer 5/10 The Magic Cycle Connecting Fundamental Research and Product Innovation Kantrowitz challenges Matias's 'magic cycle' model by noting that quarter-to-quarter product goals can undermine long-term research focus. Matias acknowledges the tension and outlines how Google balances research priorities.10:41–16:22 · Alex as informed peer 4/10 Distinguishing Breakthroughs from Innovation and Long-Term Horizons When Kantrowitz asks about detached long-term research, Matias rejects the premise, asserting that no effective research is truly detached from real-world possibilities, citing Transformers and geospatial models.16:22–20:21 · Alex as informed peer 5/10 Future AI Breakthroughs: Algorithmic Innovations Versus Scaling Compute Kantrowitz forces a direct choice between algorithmic innovation and scaling compute. Matias breaks down why both scaling and architectural/reasoning advances remain necessary.20:22–23:01 · Alex as informed peer 4/10 Motivating Researchers and Google's Multi-Disciplinary Full-Stack Approach Kantrowitz asks how Google keeps researchers motivated on non-LLM domains amid industry hype. Matias details the intrinsic motivations of scientists and Google's multi-disciplinary full-stack environment.23:01–26:37 · Alex as informed peer 5/10 AI as an Amplifier of Human Ingenuity and the Future of Researchers Kantrowitz queries whether automated AI discovery tools will eliminate researcher jobs. Matias uses AlphaFold and AI co-scientists to demonstrate why AI expands the questions scientists can tackle rather than reducing headcount.3:10–6:54 · Guest teaching 4/10 Quantum Computing Breakthroughs and Long-Term Real-World Impact Kantrowitz asks an informed question comparing recent headline breakthroughs with the recurring 5-to-10-year timeline disconnect. Matias explains Nature benchmark verifications and sets expectations for practical application within five years.6:55–10:40 · Guest teaching 4/10 The Magic Cycle Connecting Fundamental Research and Product Innovation Kantrowitz challenges Matias's 'magic cycle' model by noting that quarter-to-quarter product goals can undermine long-term research focus. Matias acknowledges the tension and outlines how Google balances research priorities.10:41–16:22 · Guest teaching 5/10 Distinguishing Breakthroughs from Innovation and Long-Term Horizons When Kantrowitz asks about detached long-term research, Matias rejects the premise, asserting that no effective research is truly detached from real-world possibilities, citing Transformers and geospatial models.16:22–20:21 · Guest teaching 4/10 Future AI Breakthroughs: Algorithmic Innovations Versus Scaling Compute Kantrowitz forces a direct choice between algorithmic innovation and scaling compute. Matias breaks down why both scaling and architectural/reasoning advances remain necessary.20:22–23:01 · Guest teaching 3/10 Motivating Researchers and Google's Multi-Disciplinary Full-Stack Approach Kantrowitz asks how Google keeps researchers motivated on non-LLM domains amid industry hype. Matias details the intrinsic motivations of scientists and Google's multi-disciplinary full-stack environment.23:01–26:37 · Guest teaching 5/10 AI as an Amplifier of Human Ingenuity and the Future of Researchers Kantrowitz queries whether automated AI discovery tools will eliminate researcher jobs. Matias uses AlphaFold and AI co-scientists to demonstrate why AI expands the questions scientists can tackle rather than reducing headcount.3:10–6:54 · Guest disagreement 1/10 Quantum Computing Breakthroughs and Long-Term Real-World Impact Kantrowitz asks an informed question comparing recent headline breakthroughs with the recurring 5-to-10-year timeline disconnect. Matias explains Nature benchmark verifications and sets expectations for practical application within five years.6:55–10:40 · Guest disagreement 1/10 The Magic Cycle Connecting Fundamental Research and Product Innovation Kantrowitz challenges Matias's 'magic cycle' model by noting that quarter-to-quarter product goals can undermine long-term research focus. Matias acknowledges the tension and outlines how Google balances research priorities.10:41–16:22 · Guest disagreement 2/10 Distinguishing Breakthroughs from Innovation and Long-Term Horizons When Kantrowitz asks about detached long-term research, Matias rejects the premise, asserting that no effective research is truly detached from real-world possibilities, citing Transformers and geospatial models.16:22–20:21 · Guest disagreement 1/10 Future AI Breakthroughs: Algorithmic Innovations Versus Scaling Compute Kantrowitz forces a direct choice between algorithmic innovation and scaling compute. Matias breaks down why both scaling and architectural/reasoning advances remain necessary.20:22–23:01 · Guest disagreement 1/10 Motivating Researchers and Google's Multi-Disciplinary Full-Stack Approach Kantrowitz asks how Google keeps researchers motivated on non-LLM domains amid industry hype. Matias details the intrinsic motivations of scientists and Google's multi-disciplinary full-stack environment.23:01–26:37 · Guest disagreement 1/10 AI as an Amplifier of Human Ingenuity and the Future of Researchers Kantrowitz queries whether automated AI discovery tools will eliminate researcher jobs. Matias uses AlphaFold and AI co-scientists to demonstrate why AI expands the questions scientists can tackle rather than reducing headcount.3:10–6:54 · Alex pushing back 3/10 Quantum Computing Breakthroughs and Long-Term Real-World Impact Kantrowitz asks an informed question comparing recent headline breakthroughs with the recurring 5-to-10-year timeline disconnect. Matias explains Nature benchmark verifications and sets expectations for practical application within five years.6:55–10:40 · Alex pushing back 5/10 The Magic Cycle Connecting Fundamental Research and Product Innovation Kantrowitz challenges Matias's 'magic cycle' model by noting that quarter-to-quarter product goals can undermine long-term research focus. Matias acknowledges the tension and outlines how Google balances research priorities.10:41–16:22 · Alex pushing back 3/10 Distinguishing Breakthroughs from Innovation and Long-Term Horizons When Kantrowitz asks about detached long-term research, Matias rejects the premise, asserting that no effective research is truly detached from real-world possibilities, citing Transformers and geospatial models.16:22–20:21 · Alex pushing back 3/10 Future AI Breakthroughs: Algorithmic Innovations Versus Scaling Compute Kantrowitz forces a direct choice between algorithmic innovation and scaling compute. Matias breaks down why both scaling and architectural/reasoning advances remain necessary.20:22–23:01 · Alex pushing back 3/10 Motivating Researchers and Google's Multi-Disciplinary Full-Stack Approach Kantrowitz asks how Google keeps researchers motivated on non-LLM domains amid industry hype. Matias details the intrinsic motivations of scientists and Google's multi-disciplinary full-stack environment.23:01–26:37 · Alex pushing back 3/10 AI as an Amplifier of Human Ingenuity and the Future of Researchers Kantrowitz queries whether automated AI discovery tools will eliminate researcher jobs. Matias uses AlphaFold and AI co-scientists to demonstrate why AI expands the questions scientists can tackle rather than reducing headcount.

speaking balance: gold is Alex, purple is the guest (3 minute bins)

0:00 · Alex 28.7% · guest 71.3%0:00 · Alex 28.7% · guest 71.3%3:00 · Alex 22.3% · guest 77.7%3:00 · Alex 22.3% · guest 77.7%6:00 · Alex 4.4% · guest 95.6%6:00 · Alex 4.4% · guest 95.6%9:00 · Alex 21.9% · guest 78.1%9:00 · Alex 21.9% · guest 78.1%12:00 · Alex 3% · guest 97%12:00 · Alex 3% · guest 97%15:00 · Alex 8% · guest 92%15:00 · Alex 8% · guest 92%18:00 · Alex 2.9% · guest 97.1%18:00 · Alex 2.9% · guest 97.1%21:00 · Alex 16.3% · guest 83.7%21:00 · Alex 16.3% · guest 83.7%24:00 · Alex 1.1% · guest 98.9%24:00 · Alex 1.1% · guest 98.9%
Sharpest disagreement ▶ 12:47 Guest rejects premise of detached research

Matias immediately counters the host's framing with 'First, no research is detached,' asserting that all meaningful basic research is tied to the art of the possible.

Hardest push from Alex ▶ 9:09 Host challenges quarterly product incentives corrupting research

Kantrowitz directly challenges Matias's synergy narrative by pointing out that product teams prioritize quarterly growth metrics over long-term research horizons.

Biggest teaching moment ▶ 23:30 Guest details research expansion economics via AlphaFold

Matias clearly educates the host on research dynamics, explaining that automating protein folding did not reduce protein researchers but allowed them to pursue larger scientific questions.

Alex holds their own ▶ 2:09 Host demonstrates pre-interview reading on cancer cell mechanisms

Kantrowitz showcases independent domain preparation by detailing the exact biological mechanism where the AI model prompted cancer cells to signal the immune system.

the scores for every segment, with the reasoning behind each
ChapterTopicAlex as informed peerGuest teachingGuest disagreementAlex pushing backWhy
Quantum Computing Breakthroughs and Long-Term Real-World Impact 4413 Kantrowitz asks an informed question comparing recent headline breakthroughs with the recurring 5-to-10-year timeline disconnect. Matias explains Nature benchmark verifications and sets expectations for practical application within five years.
The Magic Cycle Connecting Fundamental Research and Product Innovation 5415 Kantrowitz challenges Matias's 'magic cycle' model by noting that quarter-to-quarter product goals can undermine long-term research focus. Matias acknowledges the tension and outlines how Google balances research priorities.
Distinguishing Breakthroughs from Innovation and Long-Term Horizons 4523 When Kantrowitz asks about detached long-term research, Matias rejects the premise, asserting that no effective research is truly detached from real-world possibilities, citing Transformers and geospatial models.
Future AI Breakthroughs: Algorithmic Innovations Versus Scaling Compute 5413 Kantrowitz forces a direct choice between algorithmic innovation and scaling compute. Matias breaks down why both scaling and architectural/reasoning advances remain necessary.
Motivating Researchers and Google's Multi-Disciplinary Full-Stack Approach 4313 Kantrowitz asks how Google keeps researchers motivated on non-LLM domains amid industry hype. Matias details the intrinsic motivations of scientists and Google's multi-disciplinary full-stack environment.
AI as an Amplifier of Human Ingenuity and the Future of Researchers 5513 Kantrowitz queries whether automated AI discovery tools will eliminate researcher jobs. Matias uses AlphaFold and AI co-scientists to demonstrate why AI expands the questions scientists can tackle rather than reducing headcount.

Statements from this episode (5)

Assertion Partly supported
Google claims first verifiable practical quantum application advantage over classical computers
“Our this announcement of yesterday is a paper in nature that that actually shows the first verifiable practical application advantage of a quantum computer over classical computer.”
Yossi Matias Oct 28, 2025 ▶ 4:51
Prediction Open · timeframe Oct 2030
Practical quantum computing applications will emerge within five years, Matias predicts
“And I'm quite optimistic that we are going to see these real life applications in the framework of about five years.”
Yossi Matias Oct 28, 2025 ▶ 5:22
Assertion Supported
Google's AI flood forecasting model covers two billion people globally
“Our work on flood forecasting started in in 2017. Now we have a global model serving two billion people in 150 countries.”
Yossi Matias Oct 28, 2025 ▶ 15:19
Prediction Not checkable as stated
Matias predicts generative AI will increase global demand for scientific researchers
“Well, we're going to need many more researchers in all disciplines.”
Yossi Matias Oct 28, 2025 ▶ 23:32
Assertion Supported
AlphaFold's release increased the total number of researchers working on proteins
“I mean, we don't have less researchers working on proteins. We have actually have many more, right?”
Yossi Matias Oct 28, 2025 ▶ 24:22
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 300 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.