The Flywheel by M13

M13
  • 0.0 (0)
  • Entrepreneurship

The Flywheel is a podcast for founders, operators and investors building through structural change. M13 partners sit down with the founders reshaping AI, enterprise software, healthcare, commerce, and consumer technology to unpack the hard calls and inflection points as they build category-defining companies. Each conversation is designed to help listeners better understand where the world is changing, what great founders see before others do, and how enduring companies are built.

Episodes

  1. 2d ago

    The MIT Founder Who Bet on the Market Nobody Wanted | Parth Shah, Polimorphic

    What if the most overlooked market in tech is the one every investor told you to run from? Polimorphic founder Parth Shah and M13 Managing Partner Latif Peracha explore how local governments are leapfrogging a decade of technology straight into AI and ask a bigger question: can AI reduce bureaucracy without reducing government? Recorded live at M13's Annual General Meeting, Latif and Parth break down why off-the-shelf models score just 16% accuracy on dense government data while Polimorphic hits 99%, how AI can reduce the “time tax” of bureaucracy, and why the competitive whiteboard stayed almost entirely white. Parth launched Polimorphic in 2021 when investors said they'd rather he "have no idea than work with local government.” They’re in 30 states today, and when governments think AI, they think Polimorphic. If you're deciding where to build, weighing whether your industry is ready to be disrupted, or studying how companies compound in slow markets, this one’s for you. ⏱️ Chapters 00:00 — Intro: the overlooked market hiding in plain sight02:13 — The market every investor told him to avoid05:01 — Government is a customer service organization first05:29 — Running the company from a desk inside town hall06:28 — The product: what Polimorphic actually does08:01 — From 16% to 99% AI accuracy09:04 — Cutting permitting from months to two weeks09:45 — The whiteboard that stayed white: where's the competition?10:21 — How Polimorphic grows: city to county to state13:14 — Pricing per resident and the $200M North Carolina market13:52 — The $1 trillion labor shift hitting local government14:26 — 30 states and the Michigan ERP partnership15:52 — Being a first-time founder17:48 — What a great investor does when things get hard🏛️ In this episode Why out-of-the-box LLMs hit just 16% accuracy on government data—legally dense, multilingual, fifth-grade reading level—and how Polimorphic gets to 99%How the “time tax” may be one of AI’s most consequential government use cases: Polimorphic cut permitting from three-plus months to two weeks by helping local officials eliminate the administrative work behind the wait.Why one county told its commission that Polimorphic delivers 30% of a department's staff value—work that would otherwise take three new hiresThe compounding growth loop: 20%+ of North Carolina's population covered, a $200M market in that one state at $5–6 per residentThe macro thesis: state and local government carries a $2.2 trillion labor budget and is set to lose 30–50% of its staff in three to five years—leaving $600B–$1T of work that has to get doneWhy Polimorphic became North Carolina's de facto AI customer service platform, plus a new exclusive partnership with the ERP provider serving 95% of Michigan governmentsAbout Polimorphic Polimorphic builds AI-powered customer service, CRM, and workflow tools that help local governments serve residents faster—from answering questions to processing permits. Learn more at https://www.polimorphic.com About M13 M13 invests early in outlier founders building at the structural shifts that create new markets. The Flywheel is where M13 talks with those founders. Learn more at https://m13.co ▶️ Subscribe Subscribe for more conversations with the founders building at the structural shifts that create new markets, and follow The Flywheel on Spotify and Apple Podcasts.

    The MIT Founder Who Bet on the Market Nobody Wanted | Parth Shah, Polimorphic

  2. 5d ago

    Dropping Out to Build 911 AI | Mike Chime, Prepared

    Why does calling 911 still mean a dispatcher typing your emergency by hand, word for word? Prepared co-founders Mike Chime, Neal Soni and Dylan Gleicher dropped out of Yale to rebuild the emergency call—using AI to transcribe, summarize, and translate in real time across roughly 80 million calls a year. Prof G Markets cohost Ed Elson sits down with Mike Chime, co-founder and CEO of Prepared, and M13 Managing Partner Karl Alomar to break down what it takes to fix an overlooked problem: why the 911 call is still stuck on decades-old technology, how Prepared now works with about 1,000 agencies and touches roughly 80 million calls a year, and why solving a real problem beats chasing the hot trend.  Mike—who left Yale on a Thiel Fellowship—explains how the company went from a dorm-room school-safety app to being acquired by Axon. Karl, who scaled DigitalOcean to a $5 billion IPO before joining M13 and wrote Prepared’s first check, explains what made him back college students a year before there was a company.  If you're a founder, operator, student or engineer weighing whether to leave a safe path and build something that matters, this one's for you. ⏱️ Chapters  00:00 — Why this conversation matters more than ever04:05 — What is Prepared? Fixing the 911 call with AI 06:20 — How Karl found Mike at Yale 09:26 — What made a college kid worth investing in 11:56 — From a dorm-room school-safety app to 911 14:59 — The Thiel Fellowship and burning the boats 16:31 — Should you drop out of college to start a company? 20:02 — "The 911 thing": surviving the skepticism 21:16 — Chasing the hot thing vs. solving a real problem 25:10 — What scaling taught Mike about management 28:12 — Has AI changed how you manage people? 30:43 — Taking risks and betting on yourself 32:21 — Advice for anyone on the fence🚨 In this episode Why the 911 call is still stuck on decades-old tech—dispatchers typing what they hear, and callers who don't speak English waiting five or six minutes for a translatorHow Prepared uses AI to transcribe, summarize, and translate emergencies in real time, now across about 1,000 agencies and roughly 80 million calls a yearThe founder path most people romanticize: dropping out on a Thiel Fellowship, seven years of "the 911 thing," and the skepticism before Axon acquired the companyWhy solving a real problem beats chasing the hot trend—and how that shows up in what you build and who you hireFounder mode, management, and why AI changes org structure but not the need for people who believe in the missionAbout Prepared  Prepared helps 911 centers handle emergencies with AI—transcribing and summarizing calls in real time, surfacing what responders need, and translating for callers who don't speak English. Learn more at https://www.prepared911.com About M13  M13 invests early in outlier founders building at the structural shifts that create new markets. The Flywheel is where M13 talks with those founders. Learn more at https://m13.co ▶️ Subscribe for more conversations with the founders building at the structural shifts that create new markets, and follow The Flywheel on Spotify and Apple Podcasts.

    Dropping Out to Build 911 AI | Mike Chime, Prepared

  3. Aug 6

    The Future of Cross-Border Payments | OpenFX CEO Prabhakar Reddy

    What if moving $100 million across borders took 60 minutes instead of five days? OpenFX founder Prabhakar Reddy is rebuilding the world's cross-border payment rails to make real-time the default. On this episode of The Flywheel, recorded live at M13's annual meeting in Montana, M13 Managing Partner Latif Peracha sits down with Prabhakar Reddy, co-founder and CEO of OpenFX. OpenFX is Prabhakar's fifth company. He's been a founder since 19, spent a stint as a VC at Accel, and previously built FalconX into a profitable $8 billion company where he's still a large shareholder. Now he's taking on a $2-trillion-a-day problem: moving money across borders in real time. Prabhakar breaks down how OpenFX settles 98% of transactions in under 60 minutes, why the real competition is legacy banks and not other fintechs, and how the company grew more than 20x in a year while staying profitable with roughly 100 employees. In this episode: Why cross-border money still takes three to five days—and what that delay costs when a currency drops 5% in a weekHow OpenFX uses stablecoins as invisible infrastructure to move money in minutes, without clients ever touching cryptoThe growth story: from $2 billion to roughly $50 billion in volume in about a year, and the internal target of $1 trillionWhy the durable moat is solving the whole puzzle—compliance, licensing, market making, collection, and payouts—not just one leg of the tradePrabhakar's biggest worry as a CEO: not security or regulation, but staying fast enough to avoid becoming "another slow company"Chapters:  00:00 — The longest journey to Montana: meet Prabhakar Reddy 01:40 — Fifth company, an $8B exit, and why he keeps going 02:55 — Why cross-border money movement is broken, and what it costs 04:40 — Moving money in minutes: stablecoins as invisible rails 06:50 — 20x in a year: the growth story and the Wise comparison 09:25 — Picking investors as a former VC, and why M13 got the exception 12:25 — The moat: why banks, not fintechs, are the real competition 16:00 — The Amazon-esque model: wholesale FX and 90% cheaper rails 17:55 — What stablecoin regulation could (and couldn't) break 20:05 — From security threats to the fear of becoming a slow company 22:40 — Operating leverage: keeping opex flat while volume explodesFollow The Flywheel on Spotify or Apple Podcasts for more conversations with the founders building at the structural shifts that create new markets. Learn more about OpenFX at openfx.com and M13 at m13.co.

    The Future of Cross-Border Payments | OpenFX CEO Prabhakar Reddy

  4. Aug 6

    Voice AI's Rising Bar: Rime CEO Lily Clifford on Conversation Quality

    What makes an AI voice worth talking to? Rime founder Lily Clifford thinks the next breakthrough in voice AI isn't a more natural-sounding voice—it's a conversation you actually want to keep having. On this episode of The Flywheel, M13 Partner Morgan Blumberg sits down with Lily Clifford, co-founder and CEO of Rime. Before Rime, Lily was earning a linguistics PhD at Stanford, researching the acoustic physics of how humans produce speech. Lily’s since built Rime into voice AI that powers roughly 100 million phone interactions a month across healthcare and financial services, and has a clear, differentiated view of where voice goes next. Lily breaks down why the next wave isn't a smoother-sounding voice, but conversations people don't want to hang up on and what that means for anyone building with voice right now. In this episode: Why "quality of interaction" is a far higher bar than natural-sounding audio—and why there's no margin for error when someone expects a humanWhat an independent study of roughly 100,000 calls revealed about why people stay on the phone longer with Rime than with Eleven Labs or GoogleLily's take on why you can't actually build on today's most impressive frontier voice modelsWhy enterprises will start building their own voice stacks—and the building blocks that are still missingWhy a linguistics-first team treats "impossible" problems (like how to pronounce the word "live") as the whole pointChapters:  00:00 — Why "natural" isn't the bar for voice AI 01:00 — 25th Street Recording: where Rime made its first recording 04:00 — Quality of interaction vs. quality of voice 08:29 — The Miravoice study: ~100,000 calls and the 10-second effect 12:12 — Build vs. buy: why frontier voice models aren't buildable 16:13 — The building-blocks problem and the enterprise "game of telephone" 21:40 — Will enterprises build voice in-house? The SaaS parallel 26:55 — A linguistics-first team, and leaving a Stanford PhD 36:23 — Why chase an "impossible" problem 41:00 — Language, vowels, and what makes us human 46:15 — What keeps Lily goingFollow The Flywheel on Spotify or Apple Podcasts for more conversations with the founders building at AI's frontier.  Learn more about Rime at rime.ai and M13 at m13.co.

    Voice AI's Rising Bar: Rime CEO Lily Clifford on Conversation Quality

  5. Aug 6

    What Comes After LLMs? Sam Pasupalak, Skyfall AI

    What comes after large language models? Skyfall AI founder Sam Pasupalak thinks the next breakthrough is AI that can run a business autonomously.  On this episode of The Flywheel, M13 Partner Morgan Blumberg sits down with Sam Pasupalak, co-founder and CEO of Skyfall AI. Before Skyfall, Sam co-founded Maluuba, one of the first deep-learning labs for natural language understanding, acquired by Microsoft in 2017. He's spent 15+ years at the frontier of AI and he has a clear, differentiated view of where it goes next. If you're a founder, operator, or engineer trying to understand the future of AI after the LLM era, this is your map. Sam breaks down why the next wave isn't smarter co-pilots, but autonomous businesses—and what that means for anyone building right now. In this episode: What "world models" and continual learning are, and why Sam thinks they are the real breakthroughSam’s hot take on why so many AI "wrapper" companies will struggle to surviveHow early we actually are in the AI cycle (think Yahoo and AOL, before Google and Amazon)Why 20 million software engineers and a billion white-collar workers signal a shift bigger than the Industrial RevolutionThe founder traits it takes to build through a platform shift—and the sacrifice Sam says it demandsChapters: 00:00 — The future isn't AI co-pilots 00:40 — From Maluuba to Skyfall: who is Sam Pasupalak? 07:45 — Why Skyfall takes a different path than the LLM wrappers 09:50 — World models and continual learning, explained 19:17 — The coffee-shop test: why an LLM can't run a business 22:47 — The hot take: why AI wrapper companies will struggle 24:21 — How early we are: the Yahoos and AOLs vs. the Googles and Amazons 38:13 — Founder traits for this era: build a research-first company 39:46 — Hot takes on OpenAI, Anthropic, Altman and Amodei 48:43 — Being both a warrior and a monk Follow The Flywheel on Spotify or Apple Podcasts for more conversations with the founders building at AI's frontier. Learn more about Skyfall AI at skyfall.ai and M13 at m13.co.

    What Comes After LLMs? Sam Pasupalak, Skyfall AI
  • 5 Episodes

About

The Flywheel is a podcast for founders, operators and investors building through structural change. M13 partners sit down with the founders reshaping AI, enterprise software, healthcare, commerce, and consumer technology to unpack the hard calls and inflection points as they build category-defining companies. Each conversation is designed to help listeners better understand where the world is changing, what great founders see before others do, and how enduring companies are built.

Information

  • Creator
    M13
  • Years Active
    2026
  • Episodes
    5
  • Rating
    Clean
  • Copyright
    © 2026 The Flywheel by M13
  • Show Website
    The Flywheel by M13