Passion Play Profit

Peter Yuan Lu

Nothing feels better in life than discovering your passion, playing the game of business, and creating an lasting impact. Join Peter Lu as he interviews the world's leading entrepreneurs to extract hyper-practical insights to excel in business and life. Whether you’re looking to start your own company, need help motivating your team, or want to catch up on events in Silicon Valley, Wall Street, and beyond, your journey should begin here.

  1. Oct 26

    #56 - Why Every Founder Is Secretly Running From Boredom, Maskless Lithography, and AI Speed with Imad Agha of Nabu Optical Systems

    In this conversation, Imad Agha, founder and CEO of Nabu Optical Systems and a professor of quantum optics and nano optomechanics at the University of Dayton, shares how research on optical memories turned into a maskless lithography startup, why AI is compressing the path from idea to implementation, and why founders are always running from boredom. Imad studied engineering in Beirut, did grad school in the US and a postdoc in France, and spent 11 years as a professor before spinning out Nabu with students from the lab. A phase change material the lab was writing with a laser turned out to change its solubility too, which is the basis of lithography. Paired with flat lenses, it became a tool that builds chips without physical masks. We discuss: - Why physics and optics answer the questions engineers skip - Teaching as a two way street - AI democratizing knowledge, and why pure design businesses should rethink - Why professors should spin out transformative ideas - From optical memories to maskless lithography - What accelerators teach you about market size - Reaching enterprise customers through champions - Why the future of chips won't be incremental - Designing a wafer scale lens system in weeks with AI agents - Vertical integration and a 90% cost cut - Why founders fear boredom Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  2. Oct 20

    #55 - Building Data Centers In 110 Days, From Bitcoin Mining To AI, and the Power Bottleneck with Sivan Sud of Cormint

    In this conversation, Sivan Sud, Executive Vice President at Cormint, shares how Cormint built a data center phase in 110 days instead of the usual 12 to 18 months, how moving construction into the factory cut build costs to a fifth, and what it takes to move a West Texas Bitcoin mining site to AI. Sivan studied mechanical engineering at Cornell and co-founded Ecosystems, a hardware startup out of Cornell's eLab. Sivan then started doing contract work for Cormint, designed a 200 kilowatt immersion cooled module that the company multiplied about 200 times, and has since designed, built and commissioned more than 150 megawatts of modular data centers. We discuss: - Lessons from a B2C hardware startup, and agreeing on goals early - Designing immersion cooling for one of the hottest places to build - Building the sixth phase in 110 days - Cutting build costs to one fifth by moving work into the factory - Why Bitcoin miners are pivoting to AI - First principles: compute 20 feet from the transformer - The real AI data center bottlenecks: power and skilled labor - Why West Texas - Data centers at sea and in space, and the radiation problem - The supply chain as a flow of dependencies - Why young people should get hands on experience - Density, DC power and bring your own generation - Why AI will look like Bitcoin mining Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  3. Oct 14

    #54 - Why AI Rewards Generalists, How VCs Really Win Deals, and Building Seattle's AI Scene with Sri Chandrasekar of AI House

    In this conversation, Sri Chandrasekar, Managing Director at AI House and former Partner at Point72 Ventures, shares why AI rewards broadly skilled people over specialists, what investors can really assess at each stage of a company, and why he believes Seattle can build the next trillion dollar company. Sri was born in India, grew up in Columbus, Ohio, and studied engineering at Cornell. He graduated into the dot com bust, built communications technology at BAE Systems after 9/11, then invested for the intelligence community at In-Q-Tel, where a bet on machine learning in 2014 changed the direction of his career. He spent nearly nine years helping build the venture business at Point72, and today writes first checks into applied AI founders at AI House in Seattle. We discuss: - Mistiming the market twice, and taking what the world offers - In-Q-Tel: matching startups to the missions of three letter agencies - How ad targeting and early deep learning pulled him into AI - The specialist is dead: why AI rewards Renaissance people - Building a venture business from scratch at Point72 - What you can actually assess at seed, Series A, Series B and beyond - Why the best VCs win on salesmanship, not assessment - What founders really pick investors for - Mentors who filled his gaps - Why founders should learn from other founders right now - AI House and the Seattle founder community - What he wants to back: hardware plus AI, and AI enabled service businesses Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  4. Oct 13

    #53 - Solving Real Problems, Talking To Customers, and Building With AI Coding Tools with Nicole Hao of Janet AI

    In this conversation, Nicole Hao, Founding Engineer at Janet AI (YC S25) and founder of NeuroAbility, shares why most founders are solving a problem that doesn't exist, how to talk to customers so you learn what they actually need, and how she gets the most out of AI coding tools. Nicole came from a family where neither parent went to college, studied math, computer science and neuroscience at Cornell and Harvard, and started out as an astrophysics researcher. Meeting a visually impaired math PhD student who still paid $100 an hour for someone to take notes for him showed her a problem research alone could not solve. She took it through Cornell's eLab, built Inxite AI, and turned it into NeuroAbility to widen educational access. Today she is a founding engineer at Janet AI, an AI native product management system for engineering and product teams. We discuss: - From astrophysics research to building things that change people's lives - The eLab Demo Day moment that made her want to be an entrepreneur - Why Inxite AI became NeuroAbility - Talking to 200 people: students, professors and administrators - How to check whether your problem is real - Ask about their struggles, not whether they like your idea - The customer is your boss, and they can fire you anytime - Janet AI: agents that plan, track and close tickets for you - How she uses Claude and Cursor, and why LLMs can't see the vision in your head - Rules files, clear instructions and real context: trash in, trash out - Passion, focus and chasing flow state Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  5. Oct 8

    #52 - Building A Brain ASIC With A Team Of Three, Robot Fleets, and Breaking Into Deep Tech with Jeff Ciesielski of LeafLabs

    In this conversation, Jeff Ciesielski, Director of Research at LeafLabs and stealth founder, shares how a team of three built a 2,000 channel brain recording ASIC that most people called impossible, how to learn anything hard, and how to break into deep tech from a white collar job. Jeff grew up on a farm in rural Indiana, got his first engineering job at 17 porting Fortran to C# at a steel mill, and moved to Seattle with nothing left in the bank. He worked as a mechanic and in a sandwich shop, automated his own job at a Bering Sea fishing company, subcontracted on the McLaren MP4-12C supercar, scaled robotics and controls at KeyMe from 125 to 5,000 machines, and now leads unusual research projects in neuroscience, medical devices and compute at LeafLabs. We discuss: - How to learn anything hard: chip off a piece, eat what you can, sleep on it - Knowing when it is okay to wing it, and when you need hard answers - Why you should never be the smartest person in the room - Building a 2,000 channel electrophysiology ASIC with a team of three - Working with TSMC through middlemen, shuttle runs and 180 nm process nodes - Etching silicon probes the width of a human hair - Real time process monitoring for robot fleets and AI agents - Breaking into deep tech: learn a trade that needs things built for it - Doing work that matters long after anyone remembers your name Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  6. Oct 5

    #51 - Why Stitching AI Tools Together Breaks Your Workflow, Getting Into YC, and the One-Person Company with Richard Song of Epsilla

    In this conversation, Renchu (Richard) Song, Co-Founder and CEO of Epsilla (YC S23), explains why stitching AI tools together breaks your workflow, how to get into YC, and why teams of three to five people will soon run businesses doing tens of millions in revenue. Richard dreamed of building a software business since he was ten, studied computer science, and co-founded an indoor navigation startup at Cornell. He joined TigerGraph before its Series A, spent a year at Meta on WhatsApp privacy infrastructure, and returned to lead a 30-person team at TigerGraph before starting Epsilla, which went from vector database to AI agents to a business operating system for the one-person company. We discuss: - Startup or big company out of school? Maximize for learning, not money - Why judgment and taste are the skills that matter in the AI age - Product manager, engineer and designer merging into one builder - Customer discovery with The Mom Test - How to get into YC: founder market fit and a story with clear logic - Harness Router: building on Claude Code and Codex instead of your own harness - Open source harnesses, model choice, and why you need your own evals - The business operating system for 3 to 5 person companies Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  7. Sep 30

    #50 - How the Memory Market Shapes AI, the US-China Chip Race, and the Memory Cycle with Vis Valluri of SK hynix

    In this conversation, Vis Valluri, Senior Director of Industry Intelligence at SK hynix, explains how the memory market shapes AI, how close China really is in the chip race, and what three decades in semiconductors taught him about one of the most cyclical industries in tech. Vis grew up in India, studied mechanical engineering at IIT, and earned his master's at SUNY Binghamton on an IBM-funded research assistantship in electronic packaging. He spent seven years as a packaging engineer at AMD, covered semiconductors and memory in equity research on Wall Street, sized the NAND market at SanDisk, and now covers every end market for DRAM and NAND at SK hynix. We discuss: - Why "there is no silicon in Silicon Valley" anymore - The US-China AI arms race and why whoever wins controls the ecosystem - Embargoes, EUV, and how close China really is on chips - From 50% margins to negative in two quarters: the memory cycle - HBM, chain of thought, agentic AI and the new memory demand drivers - Why memory looks tight through 2027 - Forecasting demand: bits per box, attach rates and hyperscaler CapEx - Advice for students: try 10 things, go deep on one or two, and talk to people Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

  8. Sep 28

    #49 - The PM Trap Every Engineer Falls Into, and Why AI Agents Are Replacing Dashboards with Anita Vithayathil of Arch Systems

    In this conversation, Anita Vithayathil, Director of Product Management at Arch Systems, shares the trap every engineer falls into when they move into product, how she learns what customers really need by spending weeks on factory floors, and why AI agents are replacing dashboards in manufacturing. Anita grew up in India, moved to the US in high school and went to four different high schools, studied electrical engineering, and started as a software engineer at Capital One before moving into product management. Today she leads product at Arch Systems, an analytics and AI platform for manufacturers. We discuss: - How to make the most of college: projects, co-ops and internships - Standing out as an engineer in the age of AI - The PM trap: engineers want to solve, product people ask why - You are not the user, even when you build for engineers - Living the customer's workflow and asking about specific moments - Data, ontology and why partners beat competitors on the factory floor - AI that sees, investigates and acts, and why a 1% scrap cut is a board-level conversation - What drives her: numbers going up and great people Watch the full video episode on YouTube: https://www.youtube.com/@PassionPlayProfit

See All (56)

Ratings & Reviews

5
out of 5
2 Ratings

About

Nothing feels better in life than discovering your passion, playing the game of business, and creating an lasting impact. Join Peter Lu as he interviews the world's leading entrepreneurs to extract hyper-practical insights to excel in business and life. Whether you’re looking to start your own company, need help motivating your team, or want to catch up on events in Silicon Valley, Wall Street, and beyond, your journey should begin here.

Information

  • Creator
    Peter Yuan Lu
  • Years Active
    2022 - 2026
  • Episodes
    56
  • Rating
    Clean
  • Copyright
    © Peter Yuan Lu
  • Show Website
    Passion Play Profit