Jul 13, 2026 · 48m · product-market-fit
Tries 1,000 ideas & kills most—how X built Waymo & Google Brain | X (Alphabet) · PMF Show
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
Astro Teller, Captain of Moonshots at X (Alphabet), shares the operational frameworks, psychological strategies, and de-risking methodologies that enable X to evaluate thousands of radical ideas and build generation-defining ventures like Waymo and Google Brain.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Pablo holds 16.6% of the talking time here. How this is scored →
Pablo as informed peer 3.1 Guest teaching 6.3 Guest disagreement 2.5 Pablo pushing back 1.5
speaking balance: gold is Pablo, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 7:33 Rejecting blind conviction and single-bet founders
Astro bluntly warns prospective entrepreneurs not to join X if they have a single pet idea they unconditionally believe in, comparing that mindset to gambling on roulette in Vegas.
Hardest push from Pablo ▶ 11:22 Challenging the premise of staffing efficiencyPablo presses Astro on whether spreading people across multiple projects is simply a cost-saving measure rather than a deliberate strategic choice.
Biggest teaching moment ▶ 26:48 Training the monkey before building the pedestalAstro uses the parable of training a monkey to recite Shakespeare on a pedestal to demonstrate why building easy infrastructure first creates a deceptive illusion of progress.
Pablo holds their own ▶ 15:24 Framing research mode vs builder modePablo demonstrates his own domain expertise by analyzing how founders fail when they prematurely shift into builder or sales mode instead of disciplined research mode.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Pablo as informed peer | Guest teaching | Guest disagreement | Pablo pushing back | Why |
|---|---|---|---|---|---|---|
| Astro Teller's Background and the Origin of X | 3 | 5 | 2 | 1 | Pablo sets the stage by contrasting traditional venture-backed single-focus founders with X's institutional model. Astro explains the charter and structure of X as a 21st-century Bell Labs trying 1,000 to 1,500 ideas per decade. | |
| The Card-Counter Approach to Innovation | 3 | 6 | 3 | 1 | Pablo shares his initial struggles with idea generation as a former founder. Astro reframes the pursuit of innovation from gambling on lucky bets to behaving like card counters who systemically test and discard flawed ideas without emotional attachment. | |
| Structuring Teams and the Early Idea Funnel | 4 | 6 | 2 | 2 | Pablo inquires whether having part-time roles on early projects is driven by cost, prompting Astro to explain it is entirely about psychological detachment. Astro emphasizes stripping away founder ego and not tracking who originally proposed an idea. | |
| Defining Moonshots and Case Studies in Product Framing | 4 | 7 | 3 | 2 | Pablo notes that founders often rush into building rather than methodical research. Astro dismantles the obsession with frantic progress using the lost-in-a-rowboat metaphor and walks through Waymo's evolution from car manufacturing to becoming the safest driver. | |
| The Monkey and the Pedestal De-risking Framework | 3 | 7 | 2 | 1 | Pablo asks for practical frameworks to identify testable hypotheses. Astro details the famous monkey and pedestal metaphor, explaining how attacking the hardest risk first avoids wasteful work on easily solvable components. | |
| Project Graduation and Prioritizing Learning over Progress | 2 | 6 | 2 | 1 | Pablo asks about subsequent phases after de-risking, and Astro clarifies that the process remains continuous learning rather than traditional progress tracking. Astro rejects conventional product-market fit in favor of broader system transformation. | |
| Kill Criteria and Fostering Creative In-Person Culture | 3 | 6 | 3 | 2 | Pablo asks how to prevent zombie projects from lingering for years. Astro outlines pre-committed kill criteria and shares how high-trust in-person collaboration enables teams to kill their own initiatives, such as the fusion project. | |
| Reimagining Product-Market Fit: Moonshot Market Fit | 3 | 7 | 3 | 2 | Pablo probes what true customer pull looks like, and Astro critiques conventional product-market fit as potentially just building a ladder to the moon. Astro explains that transforming massive infrastructure, like the electric grid, requires deep ecosystem alignment rather than immediate software sales. |
Statements from this episode (21)
Teller: Creating moonshots is easy by dumping money on smart people
“Taking moonshots is really easy. Super easy. If you don't care about efficiency, you find really high energy people who think they're right and are pretty smart. You dump a bunch of money on them and you'll get some moonshots or anyway, some like fairly big co…”
Teller: X evaluates 1,000 to 1,500 ideas per decade
“And so our job is to try 1000, maybe 1500 ideas per decade. And if we get a couple Google brains or Waymo's or wings per decade, Especially if we're really efficient at throwing most of those ideas away early, then this is really good for our investor alphabet…”
Teller: X operates as the card counters of innovation
“At X, we're trying to be the card counters of innovation.”
Teller: X gives founder shares to teams whose projects graduate
“Finally, you try one and it actually gets all the way through X. You graduate. We give that team founders shares like a lot of the company goes to them and you took no risk the whole time while you were doing it.”
Teller: X evaluates 200 early-stage ideas annually with zero full-time staff
“So at the very early part of the process, we look at one to 200 ideas per year.
In enough detail that they get a code name in the very early part of the process.
No person is full time on any one of those projects.
At the very least, we would ask you to have a…”
Teller: X bans employees from discussing who originated an idea
“We've literally don't track whose idea it was.
We don't let people talk about who founded the idea.”
Astro Teller: Innovation alpha comes from efficient discovery, not generating ideas
“Ideas are trivially easy.
We have a thousand ideas a year that many of them just don't even get code names.
Ideas are not the hard part.
Our alpha comes from how efficiently we can discover which ideas are valuable.”
Teller: Tech ecosystem poorly bridges basic science to venture funding
“We're paradigm shifting system engineers. We're trying to bridge that gap between basic science and something that's like venture investable, but that's like often a five or six year gap that gets filled very badly.”
Teller: Moonshots require a global problem, sci-fi solution, and testable tech
“A moonshot for us is huge problem with the world. Science fiction sounding solution to that problem. And some technology or set of technologies that give us A sense that this is a testable hypothesis for making that science fiction sounding solution.”
Teller: Waymo shifted framing from carmaker to mobility to safest driver
“So if the first five years were Waymo thought we're going to be making and selling a car the next five years, they thought. Actually, our job is transforming mobility, just moving people from point A to point B. That's our job. But even that framing while bett…”
Teller: Google Glass was too early, but the core thesis was right
“Google glass. Turns out we were too, too early there, but increasingly it's looking like we were very right. The whole world has come back around to that. But when we made it, we were trying to figure out, we have a sense that people want to get away from thei…”
Teller: X funds initial exploration for weeks or months, never years
“We do not fund things for five years so you can go try it. We're going to fund this for like five weeks, maybe five months, which is easy for us because you're part of X, right?”
Teller: De-risk the hardest element before building supporting infrastructure
“In this extreme situation, we can see that if you built the pedestal, you're half done, but you're really zero percent done at burning down the risk.
[1649] Pablo Srugo: That's right.
[1650] Astro Teller: The point of a testable hypothesis, especially at X, …”
Teller: X shut down its nuclear fusion project after 1,000 plasma shots
“We closed this down about a year and a half ago. We had a fusion project. We ran it. We did build something the size of a house. We did almost a thousand plasma shots. We convinced ourself it was not going to change the world. And we closed the project down”
Teller: Graduating X projects typically spin out with 10 to 50 employees
“If you do that for six years and you've burned down a lot of the risk, if we're actually at the place where like, well, there's a sizable chance this actually could be a Waymo someday. Then it graduates from X and you know, often this is sort of on the order o…”
Astro Teller: Standard product-market fit rarely changes complex global systems
“Finding some little thing that people will buy, which in some sense is a product market fit actually doesn't have a very good chance of changing the system. If our real goal is to fix the system, then we have to figure out a much more complex set of things.”
Teller: X teams must author kill criteria when going full-time
“We insist on just called kill criteria. So if your project has moved past the investigation stage, now a few of you are going to be full-time on this thing. We're tracking the costs. It's probably not public yet. There might only be three or four of you full-t…”
Teller: Remote work hurts creativity, creates silos, and reduces hard work
“And so far I've seen no evidence that we can be at our creative best when we're remote from each other. I think people work somewhat less hard, but they also work in much more siloed ways.”
Teller: X spent five years building rapport with grid operators before launching moonshot
“We spent five years doing things that would help us deeply understand their problems so that we could start helping them. And the first couple of things we did weren't even the moonshot. It was just getting a rapport going so that we could, with the grid opera…”