Jun 29, 2023 · 59m · big-technology
Google's "X" Experimental Unit CEO Astro Teller — The Incredible Promise Of AI & Its Dangers
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 a live interview with Big Technology podcast host Alex Kantrowitz, Alphabet X CEO Astro Teller demystifies artificial intelligence, unpacks the moonshot innovation methodology, and discusses deep-tech solutions for global challenges like climate change.
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 22.9% of the talking time here. How this is scored →
Alex as informed peer 4.7 Guest teaching 5.9 Guest disagreement 2.9 Alex pushing back 3.1
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 7:24 Astro rejects pausing knowledge discovery
Astro forcefully rejects the premise of curbing research, asserting that humanity should not infantilize itself by preventing the discovery of new knowledge.
Hardest push from Alex ▶ 33:31 Alex presses on why X didn't ship ChatGPT directlyAlex challenges Astro's explanation of project graduation, arguing X had the resources to launch a ChatGPT competitor independently instead of folding it into Google.
Biggest teaching moment ▶ 47:40 Astro explains moonshot compost and Project TaaraAstro reframes Loon's apparent commercial failure into a triumph of technological reuse, detailing how optical balloon tech became the high-speed laser network Taara.
Alex holds their own ▶ 26:44 Alex probes Alphabet's vulnerability to the Innovator's DilemmaAlex applies classic business strategy theory to challenge Astro on why Alphabet's moonshot lab failed to preempt OpenAI's conversational AI breakout.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Technological Duality and Lessons from Nuclear History | 5 | 5 | 2 | 3 | Alex opens with an informed personal angle, citing Astro's grandfather Edward Teller to frame technological risk. Astro delivers an eloquent historical reframing about how nuclear fear caused society to miss out on clean energy. | |
| Demystifying Artificial Intelligence and Inventor Responsibility | 4 | 6 | 4 | 3 | Alex brings up catastrophic AI paperclip scenarios and researcher risk percentages. Astro demystifies AI as 'algebra on steroids' that has been progressing steadily for 70 years rather than a sudden step function. | |
| Responsible Innovation and Real-World Sandboxing | 4 | 5 | 4 | 3 | Astro pushes back on the monolithic definition of invention, arguing humanity shouldn't infantilize itself by stopping knowledge discovery. Alex quickly clarifies his intent, prompting Astro to explain real-world sandboxing at Waymo. | |
| Debating AI Moratoriums and Prioritizing Global Crises | 5 | 6 | 5 | 6 | Alex repeatedly presses Astro on whether there is any philosophical line where humanity should pause AI development. Astro dismisses AI extinction fears as a distant third priority behind climate change and nuclear war, citing 30 years of robotic slog. | |
| Virtualizing the Electrical Grid with Project Tapestry | 4 | 7 | 2 | 2 | Alex sets up a multi-part question on practical AI applications in energy and climate. Astro delivers an in-depth explanation of the electrical grid's complexity and how Project Tapestry uses machine learning to virtualize power systems. | |
| Generative AI and the Future of Inverse Design | 4 | 6 | 3 | 2 | Alex asks whether generative AI is overhyped compared to foundational ML. Astro categorizes consumer chatbots as merely the tip of the iceberg, introducing inverse generative design across manufacturing disciplines. | |
| Democratizing Innovation and Expanding Global Moonshots | 4 | 6 | 3 | 2 | Alex asks if widespread AI democratization threatens X's competitive edge. Astro counters that more problem solvers benefit humanity, detailing Project Tidal's computer vision work for salmon aquaculture in Norway. | |
| Alphabet's Innovator's Dilemma, Google Brain, and Kodi | 7 | 4 | 2 | 6 | Alex challenges X's positioning against Clayton Christensen's Innovator's Dilemma, asking why OpenAI beat Google to market with ChatGPT. Astro acknowledges the strong setup and reveals that Google Brain and Kodi originated inside X. | |
| Prototyping Seed Crystals, Intrinsic, and Audacious Humility | 6 | 6 | 4 | 6 | Alex pushes back on graduating projects to Google, suggesting internal bureaucracy stifled ChatGPT-like commercialization. Astro explains X is a rapid prototyping lab for seed crystals rather than an operations hub for serving a billion daily users, citing Intrinsic. | |
| The Three-Circle Moonshot Framework and Fast Failure | 4 | 7 | 3 | 2 | Alex asks how X decides what ideas to fund. Astro rejects the Silicon Valley premise that anyone can predict the future, laying out X's three-circle moonshot framework and the mandate to rapidly kill bad hypotheses cheaply. | |
| The State and Nuances of Autonomous Driving at Waymo | 5 | 5 | 2 | 3 | Alex assesses the realistic timelines of autonomous driving compared to earlier hype. Astro explains the nuanced long-tail edge cases Waymo encountered around human driving habits and regulatory compliance. | |
| Moonshot Compost: From Project Loon to Project Taara | 5 | 7 | 3 | 3 | Alex asks how Project Loon lost ground to Starlink. Astro clarifies Loon's unique business model hurdles and shares the story of 'moonshot compost' where Loon's atmospheric optical tech transformed into Project Taara's ground-based laser communications. | |
| Societal Moonshots: Education, Community, and Systems Integration | 4 | 6 | 2 | 2 | Alex relays audience questions regarding societal moonshots like loneliness and community building. Astro admits education and social justice are excruciatingly difficult to solve purely via technology but explains X's role as system integrators. | |
| Techno-Economics of Climate Moonshots: Carbon Capture and Foghorn | 4 | 7 | 2 | 2 | Alex brings up carbon sequestration. Astro details the unforgiving techno-economics of climate solutions, recalling Project Foghorn which successfully synthesized seawater fuel but was terminated because cost parity could not be reached. | |
| Economic Realities, the Human Drive to Invent, and Conclusion | 5 | 5 | 2 | 2 | Alex questions whether high interest rates and tighter capital will restrict Alphabet's moonshot funding. Astro explains Alphabet's long-term orientation and summarizes humanity's intrinsic explorer spirit. |
Statements from this episode (26)
Teller: Climate change wouldn't be happening if society embraced nuclear power
“The disaster which is climate change right now would not be happening if we as a society had not let our fears about the first thing translate into an inability to use the upside of nuclear power to save us from what is now arguably the biggest problem in the …”
Teller: Artificial intelligence is just 'algebra on steroids' and math
“Artificial intelligence number one is algebra on steroids. Just so we're clear. It is a very big field of study. And when you get out of microscope and you like look down inside the computer, you cannot find the AI in there anywhere. It's just math.”
Teller: Modern commercial airplanes fly themselves almost entirely using AI
“The plane that flew you here flew itself almost the entire way using artificial intelligence, and we all rely on artificial intelligence every day, whether we realize it or not.”
Teller: Signers of AI pause letter will keep working on AI regardless
“Lots of people who signed that letter, and lots of other people in the world are going to keep working on it, no matter what letter they're signed.”
Teller: AI catastrophe ranks far behind climate change and nuclear weapons
“I think that the world is overrun with serious problems. I would put at the top of that list, climate change. Nowhere close to second place. I would put nuclear weapons, just since we use that as an example, and I would put AI doing something particularly horr…”
Teller: Extreme AI outcomes will not happen in our lifetime
“Do I believe that anything particularly extreme in our lifetime is going to happen? I don't. I'm sorry, but I've been working on robots trying to open doorknobs for like the last 30 years, and it's been a slog.”
Teller: If half the world halts AI, the other half keeps going
“And I really think if we get to that line and then some half of humanity says, okay, we're out, the other half of the humanity is just going to keep going.”
Teller: Missing digital grid maps cause 10-year delays for connecting solar power
“If you go to a system operator and say, show me the digital map of where every inverter, every transformer, every wire is in your grid, they will just say, we don't have that. And that's why it takes them 10 years. When you wonder why there are 10 year waits i…”
Teller: Completed wind and solar farms are idling due to grid bottlenecks
“Think about the tsunami of renewables that are already waiting. They've literally already been built. They're just sitting there in the dirt, wind farms and solar panels waiting because we don't have a virtualization of the grid.”
Teller: AI-guided design will transform every discipline over 30 to 40 years
“We're going to see this sort of thing play out In every discipline in the world over the next 30 or 40 years, and it will take a long time.”
Teller: Humanity derives up to $3 trillion annually from oceans
“So the humanity gets about two and a half, almost three trillion dollars a year from the oceans.”
Teller: Fish has one-eighth the carbon footprint of beef
“And because the carbon footprint of a pound of fish is one eighth the carbon footprint of a pound of beef, we are wildly better off as humanity moving to producing food through aquaculture.”
Teller: Google Brain and Transformer foundations originated inside Alphabet's X lab
“One of the early things that we did was a thing which, when we graduated back to Google, was called Google Brain. Google Brain is the origin of much of what you now think of as Bard, and so that was an example in the earlier days of X of us making something, r…”
Teller: Alphabet's X is not structured to scale products to billions
“We are not the right place to move something from, yeah, we have something special here to it's ready in a really thoughtful, responsible way to serve a billion people a day.”
Teller: Global manufacturing spending is approaching $10 trillion annually
“The world spends sort of pushing 10 trillion dollars a year making stuff.”
Teller: Seriously automating a factory currently costs $1 billion
“Right now it costs a billion dollars to seriously automate a factory.”
Teller: Innovation labs must spend capital disproving hypotheses, not proving them
“Audacious enough that we'll try almost anything, but then humble enough that we're honest right after we start out on each of these investigations that we're probably wrong and that we need to spend our money figuring out that we're wrong rather than trying to…”
Teller: No one is better than random at predicting the future
“I, one of the core axioms at X is I do not believe that anyone, certainly including myself, is any better than random at predicting the future.”
Teller: More than half of X's current projects target climate change
“More than half of what we're doing right now is in the climate change space, but not because I mandated that, but because that's what people were excited about”
Teller: Autonomous driving delays stem from human drivers' legal gray areas
“One of the things that I think caught Waymo by surprise, got us by surprise, is that there are a lot of things that human drivers do that if we want to follow the laws, we can't do. Those humans just kind of like are in the gray area a lot about how they choos…”
Teller: Project Loon worked technically but failed on commercial spectrum deals
“We built it. It was working. We were beaming LTE and five G 200 of thousands of people in multiple countries. We couldn't figure out how to get the business to close with the owners of the spectrum that we had to work through in those different companies. So w…”
Teller: Project Taara delivers 20 Gbps over 20 km at sub-0.1% fiber cost
“If you plug the internet like a fiber optic cable into one, you have 20 gigabits per second up to 20 kilometers away for less than one 1000th the cost of trenching fiber.”
Teller: Taara moves more daily data than Project Loon's entire history
“And we're rolling these out right now in mainly Africa and India. And that project is now moving more data to real customers per day than Loon moved in its entire history.”
Teller: Alphabet's X lab has tried roughly 30 education moonshots
“I'll give you another one. That's sort of like that education. Like I will consider it a failure if X doesn't eventually have a great moonshot in the education space. We've tried like 30 things. It, it's so painful.”
Teller: Cost per ton is the only metric that matters in carbon capture
“The number of dollars it costs for you to sequester a ton of carbon is like the only thing that really matters at the end of the day.”
Teller: Alphabet X shut down Project Foghorn over $15-per-gallon floor
“This project was called Foghorn. Like, that felt like a real save the world moment. It was actually working. And we could not convince ourselves we were going to get it cheaper than about 15 dollars gallon of gas equivalent. And that's just not going to save t…”