Astro has run X, Alphabet's Moonshot Factory, for 16 years. His teams created Waymo, Wing, and Google Brain—by testing over 1,000 ideas a decade and killing nearly all of them. His whole system rests on one uncomfortable rule: celebrate the people who kill their own projects.

In this episode, Astro breaks down why you should do the hard thing first, how pre-written kill criteria force intellectual honesty, and why X maximizes learning per dollar instead of progress—and never tracks whose idea it was.

Why You Should Listen

• Why working on the riskiest part first is the key to testing new ideas
• How writing kill criteria a year in advance stops you from running a zombie startup.
• Why progress doesn't matter when launching a new startup.
• How to come up with big ideas like Waymo and Wing.

Keywords startup podcast, startup podcast for founders, product market fit, finding pmf, Astro Teller, Google X, moonshot factory, Alphabet, Waymo, innovation process, kill criteria, idea validation, deep tech

Chapters

• 00:00:00 Intro
• 00:04:15 Inside Alphabet's Moonshot Factory
• 00:07:22 The Card Counters of Innovation
• 00:14:00 Learning per Dollar, Not Progress
• 00:23:04 Killing Ideas with Testable Hypotheses
• 00:26:34 Train the Monkey First
• 00:35:16 Kill Criteria: A Message to Your Future Self
• 00:41:59 The Moment of True Product Market Fit

Send me a message to let me know what you think!

Cos was the CTO of Brex, which he helped scale from 40 people to over 1,000. Before that, he led engineering at Stripe. When he finally left, he had no startup idea—just a co-founder he trusted and one rule: whatever they built had to have massive impact. Six months of brainstorming later, they landed on accounting.

In this episode, Cos breaks down why he capped the company at 21 people after raising $75M, why everyone—engineers and salespeople alike—earns the exact same salary, how he hit a 100% pilot-to-production conversion rate with top accounting firms, and the counterintuitive process he used to pick accounting over every other industry.

Why You Should Listen

• Why the only reason to join an early-stage company should be the equity, never the cash.
• How Accrual hit a 100% pilot-to-production conversion rate selling to the largest accounting firms.
• Why he raised $75M but refuses to grow past 21 people.
• How to tell real product market fit from the "just one more feature" trap.

Keywords startup podcast, startup podcast for founders, product market fit, finding pmf, AI accounting, vertical AI, enterprise sales, pilot to production, Brex, Stripe, Cosmin Nicolaescu, Accrual, lean team

Chapters

• 00:00:00 Intro
• 00:01:56 The Moment of True Product Market Fit
• 00:09:46 Collapsing 6+ Tools Into One Platform
• 00:18:07 Leaving Brex to Start From Scratch
• 00:31:41 Landing H&R Block and Armanino
• 00:34:57 A 100% Pilot-to-Production Playbook
• 00:43:26 A 21-Person Company by Design
• 00:51:38 One Salary for Everyone

Send me a message to let me know what you think!

Anurag was employee #8 at Stripe, set for life and free to do anything next. Instead he spent a year and a half hunting for a problem worth decades of his life. He chose to build a product to make it simple for developers to ship apps, going head-to-head with AWS.

In this episode, Anurag breaks down how he hit $10M ARR in 4 years with zero marketing spend, why refusing to launch a free tier was his most expensive mistake, and how putting engineers on customer support rotations quietly shaped the entire product roadmap.

Why You Should Listen

• Why you don't have real product market fit until your users sell the product for you.
• How a sub-2-minute setup turned developers into a word-of-mouth machine.
• Why skipping a free tier for years was his most expensive mistake.
• How engineers doing support on rotation built the roadmap—and 6M+ developers.

Keywords startup podcast, startup podcast for founders, product market fit, finding pmf, developer tools, cloud infrastructure, PaaS, Render, word of mouth growth, product-led growth, AI infrastructure, Stripe

Chapters

• 00:00:00 Intro
• 00:01:44 The Moment of True Product Market Fit
• 00:04:14 The #1 Driver of Word of Mouth
• 00:13:25 Why He Left Stripe to Build Render
• 00:23:48 Why AWS Would Never Build This
• 00:30:33 $10M ARR With No Marketing Spend
• 00:36:43 Engineers as the Support Team
• 00:42:19 Riding the AI Boom

Send me a message to let me know what you think!