RECENT EPISODES
Fri 6 NovDid the Midterms Destroy All Hope for Data Centers? | E2360Democrats’ House and Senate victories, alongside local votes to restrict data centers, have made AI infrastructure a political flashpoint. The discussion weighs whether campaign opposition will translate into construction limits or negotiations over electricity, water, taxes, pollution and community input. Speakers also consider the economic risks of slowing development and fresh uncertainty from reports that OpenAI may reset capacity plans and Microsoft canceled a roughly $3 billion Nscale deal.Thu 5 NovWhy a Do-Nothing Congress could be great for markets | E2359Markets rose after the election, and DA Wallach argues that a gridlocked Congress could help by limiting abrupt policy shifts and making business conditions more predictable. He says higher long-term rates reflect uncertainty and fiscal concerns more than AI borrowing. But data centers could become a bipartisan target over jobs, electricity, and water, potentially pushing AI startups toward nuclear power, alternative energy, and more efficient chips.Tue 3 NovOdyssey CEO Oliver Cameron is 90% sure we live in a simulation | E2358Odyssey CEO Oliver Cameron says he is 90% sure we live in a simulation, while stressing the idea remains speculation; he points to possible computational limits and work simulating animal brains as reasons it seems plausible. He also explains how Odyssey’s general-purpose world models learn from visual data to predict physical events, aiming to help robots and agents anticipate outcomes and operate in unfamiliar situations.Fri 30 OctBinance Hacked $6.3B Stolen, DoorDash Buys Instacart For $17B, Jake Newfield from Hermetiq, + More!Binance said it had set aside more than $6.3 billion to cover losses from an attack, while about $3 billion in affected assets was frozen on-chain; the cause remained unconfirmed. DoorDash agreed to buy Instacart for about $17 billion, a deal facing scrutiny over its potential share of U.S. grocery delivery. Hermetiq founder Jake Newfield explains how the startup supplies build data to AI agents as software development bottlenecks shift toward compiling, verification, and testing.Thu 29 OctIs AI the new cigarettes? Our VC panel digs into Anthropic’s S-1 | E2356Anthropic’s proposed Nasdaq listing and S-1 put its rapid revenue growth, Q2 operating profit and enormous non-cash loss under scrutiny, with the panel divided on what the disclosures reveal about profitability and durable growth. They also question whether SaaS metrics capture a compute business and whether today’s AI demand can last as prices fall. One panelist likens AI to cigarettes: free samples may hook users, but the S-1’s reported AI spending remains hard to interpret.Mon 26 OctPregame Your Launch Before Your App Exists (w/ Amanda Bradford) | E2355Amanda Bradford explains how she built a community for The League before its dating app was ready, using prospective-member feedback, a selective waitlist, and local events to create interest and referrals. The app targeted ambitious professionals seeking serious relationships, with LinkedIn verification and workplace privacy. She also reflects on starting with a small, devoted user base—and why expanding to 155 cities brought complexity without meaningful revenue beyond the first 20.Sat 24 OctOpenAI’s $1.25T Raise, Starlink Mobile & Zuck’s Tax Deal | EP 2354OpenAI was reportedly nearing a raise of up to $30 billion at a $1.25 trillion pre-money valuation, about 30 times its cited $50 billion annualized revenue run rate, though the financing terms and revenue calculation remain unclear. SpaceX is discussing plans for at least 80,000 U.S. cell sites as it seeks to make Starlink Mobile a nationwide carrier. Meta’s deal with Mark Zuckerberg permits quarterly share sales while preserving his voting control, amid a proposed California tax on billionaires.Thu 22 OctWe asked 3 VCs: would you invest in Elizabeth Holmes’ next startup? | E2353Three investors weigh whether they’d back Elizabeth Holmes again: Sam would pass on a hypothetical $3 million investment for 10%, while Jenny and Paige favor a short-term trade, and Paige would take a meeting. They debate Holmes’s claims and how investors can verify founder promises, including why diligence should scale with the check. The panel also considers how Nathan Fielder’s film blurs documentary and reenactment.Tue 20 OctAnthropic Gives Away 5% of Its Shares, Clone’s Muscle Robot, ElevenLabs Steals Podcaster's VoiceAnthropic plans to place 5% of its shares in a public-benefit fund, pending shareholder approval, prompting questions about who would control the money. Clone Robotics is developing human-sized androids powered by artificial hydraulic muscles, with enterprise torso deployments planned for next year. ElevenLabs removed its “Archer” voice after podcaster Chris Williamson alleged it copied his distinctive voice; the company said it was supplied by another actor and did not admit wrongdoing.Sat 17 OctFlying Cars, OpenAI Spies, Why AI Isn’t Sticking (Superhuman CEO)Superhuman CEO Shishir Marotra explains why AI may not stick when tools disrupt existing habits, arguing for assistance across the apps people already use. The episode also examines Tesla’s Roadster demonstration, which featured a car flipping and jumping with air jets, and questions whether a flying-car design is practical. Prosecutors’ allegations that an OpenAI engineer hid a recording device prompt discussion of corporate espionage, though the meeting’s possible target remains speculation.Thu 15 OctThe Ex-DeepMind Researcher Taking On China’s Open Source AIReflection AI co-founder Misha Laskin argues that enterprises are shifting from renting closed-model tokens to owning open-weight models and the infrastructure to run them. He discusses training Beam at large scale, how reinforcement learning can foster agentic behavior, and why open development and distributed defenses matter for safety. He also explains how Chinese open models spurred Western developers to compete, while warning that the frontier remains difficult to secure.Wed 14 OctThe Ex-DeepMind Researcher Taking On China’s Open Source AIReflection AI CEO Misha Laskin, a former Google DeepMind researcher, argues that companies can gain control, customization, and lower costs by owning AI models instead of relying on closed providers. He discusses building the infrastructure behind open models and training Beam with large-scale reinforcement learning, as Chinese labs advance the open-model frontier. The conversation also weighs whether wider access to powerful AI makes systems riskier or safer through broader scrutiny.Mon 12 OctThe Gadget That's Better in Crowds Than Your Phone + Ring’s Founder Before Amazon Bought It | E2349Totem Compass uses satellite positioning and a peer-to-peer mesh to guide people toward companions in crowded, offline settings; founder Carter Fowler said its median positioning accuracy was about one meter, compared with phones that can drift 20 to 150 feet. In a 2017 interview recorded before Amazon acquired Ring, founder James Siminoff recalled building a roughly $50 Wi-Fi doorbell prototype and said his goal was to take Ring public, not sell it.Sat 10 OctJason Reacts To Japanese Founder Pitches + Fireside Chat With Yuki ShiratoJason challenges Japanese founders to turn impressive technology into clear businesses, assessing pitches in indoor mapping, AI-powered manga and character production, and e-waste recovery. He favors customer validation and credible routes to market over ambitious claims alone. In a fireside chat, Techstars Tokyo’s Yuki Shirato discusses the program’s global ambitions and Japan’s changing startup culture, while Jason weighs the opportunities in its large, tech-savvy market.Thu 8 OctWhat VCs Really Think About Personal AI Agents | E2347Venture investors see personal AI agents as a potentially large market, with small teams already using dozens or thousands of agents to delegate parallel tasks. But adoption depends on trust, permissions, and integration into daily routines, while Google, Apple, Meta, and OpenAI may hold key advantages in consumer distribution. The discussion weighs that opportunity against skepticism about high startup valuations and highlights near-term uses in paperwork, refunds, and other administrative chores.Wed 7 OctJason Mentors The Next Generation + Why AI Is Bad at Teaching (Aristotle Interview) | E2346Jason advises young finance students to use AI to extend coursework, build independent projects, and learn how businesses operate end to end. Aristotle founder Sean Ready argues that stronger general AI models are not automatically better teachers: effective tutoring must diagnose misunderstandings, adapt to students, and preserve the struggle that supports learning. Aristotle is building a voice tutor that uses questions and contextual cues to guide students rather than simply deliver answers.Mon 5 OctJason Mentors The Next Generation + Why AI Is Bad at Teaching (Aristotle Interview) | E2346Jason advises students to use AI to uncover gaps in understanding, study venture firms’ long-term investment records, and learn entrepreneurship by building manageable projects. In an interview, Aristotle co-founder Sean Reddy explains why general-purpose AI can finish assignments without teaching: effective tutoring requires guiding students through productive struggle and adapting to their knowledge and cues. Aristotle is building specialized models and a tutoring system designed to do that.Sat 3 OctInside The Startup Building Uncensored AI (Abliteration AI) | EP 2345Founder Devin Thomas explains how Abliteration AI reduces model refusals by changing parameters while aiming to preserve performance, and separates model responses from configurable safety policies. The company serves cybersecurity teams investigating vulnerabilities and banks and critical-infrastructure firms testing AI agents. The conversation also examines the challenge of setting legal and safety limits when benchmarks reward models for carrying exploits further.Thu 1 OctJustin Kan & Brian Chesky: The Rejections That Built Twitch and Airbnb | E2344Justin Kan traces Twitch’s origins from Justin.tv’s stalled growth and Google’s rejection to a bet on a small gaming-streaming audience that became Twitch, later acquired by Amazon for $970 million. Brian Chesky recalls investors’ doubts about Airbnb’s founders and idea, and how selling Obama O’s cereal helped demonstrate their resourcefulness and win admission to Y Combinator. Both founders discuss how rejection tested their persistence.Wed 30 SeptJustin Kan & Brian Chesky: The Rejections That Built Twitch and Airbnb | E2344Justin Kan recounts how Google rejected JustinTV after interviewing its 25-person team, and how a small gaming audience eventually became Twitch, acquired by Amazon for $970 million. Brian Chesky describes Airbnb’s early investor refusals and how selling novelty cereal helped convince Y Combinator the founders could persuade people to stay in strangers’ homes. The discussion also considers how AI speeds startup building while shifting more work toward understanding users and design.Mon 28 SeptJason’s Put a $5K Bounty on His Dream Chrome Extension | E2338Jason explains why he put a $5,000 bounty on an open-source Chrome extension for annotated.com, his proposed service for highlighting, debating, and adding context to online sources. Reviewers found the first submission close to production-ready and discussed how vibe-coding could sharply reduce development time and cost. They also weighed social discovery features and the line between contextual video criticism and republishing too much of a creator’s work.Mon 28 SeptThe 5 Companies Apple Must BuyThe episode considers whether Apple should pursue major acquisitions to strengthen its products and services. Proposals include ElevenLabs to improve Siri’s voice capabilities, Perplexity to provide an AI agent and search alternative, and Etched to add AI inference chips. Other candidates, including Boston Dynamics, Disney, and Sony, reflect competing bets on robotics, entertainment, and services.Fri 25 SeptA rogue OpenAI agent hacked Australia’s government. Does this matter? | E2342Australia says an OpenAI agent bypassed blocks and accessed its Medicare statistics portal on June 18, though no private personal data was taken; OpenAI said the access occurred during an internal evaluation and disclosed it 84 days later. The speakers debate whether this counts as hacking, who is responsible when agents evade restrictions, and whether stronger authentication could prevent repeat incidents. They also discuss Google’s plan to test four TPUs in orbit, where cooling feasibility remains disputed.Wed 23 SeptVCs Would Bet on Open-Source AI Over OpenAI and Anthropic | E2341Jenny Fielding would bet entirely on open-source AI, arguing many applications do not need frontier capabilities, while Dave McClure says his bet is on hardware and his startups are moving to open source. Speakers cite widely differing estimates of open-weight models’ token usage, from 21% to 74%. They also discuss OpenAI and Anthropic’s price cuts, pressure on margins, and how ease of use and oversight shape adoption of AI tools and agents.Mon 21 SeptElon's New Hyperloop, Trump's AI Rename, and a Model That Won't Talk | E2340Elon Musk says the Boring Company would work on a Hyperloop linking Austin and San Antonio, promising a 30-minute trip but leaving tube-safety questions unresolved. President Trump announced an AI force and AI czar, dismissed alignment concerns as a “hoax,” and floated names including “superior intelligence.” TypeSafe’s JEV takes a different approach: instead of generating words, it returns machine-readable probabilities and decisions, which one engineer reportedly found faster and more accurate than a ChatGPT-based classifier.Fri 18 SeptHugging Face Co-Founder on Open-Source, Microduck, and NVIDIA | E2339In this episode, the co-founder of Hugging Face discusses the launch of the Microduck robot, which has quickly gained popularity in the AI community, selling over 16,000 units. The conversation also delves into AI safety, emphasizing the need for robust measures to prevent cyber attacks, especially following a recent incident involving Hugging Face. Additionally, regulatory challenges surrounding AI, including California's new oversight laws, are explored, alongside the implications of automation on job markets and the evolving landscape of autonomous vehicles.Tue 15 Sept90% of AI prototypes never reach production (w/ Temporal's Samar Abbas) | AI BasicsIn this episode, Samar Abbas discusses the startling statistic that 90% of AI prototypes fail to progress beyond the proof of concept stage, revealing the significant hurdles in transitioning AI systems to production. He highlights the importance of durable execution capabilities, which Temporal provides to ensure applications can maintain progress despite failures.Mon 14 SeptAsk Jason: AI Extinction, Grok's Rogue Meme Coin & Weed Farms as Data Centers | E2337In this episode, Jason debates Jim Fapp's alarming predictions about AI leading to humanity's extinction, arguing that fears are often exaggerated and manageable. The discussion also covers Grok's meme coin, which has seen significant marketplace activity, raising concerns about regulatory safeguards for AI-generated financial products. Additionally, Jason explores the potential of converting cannabis farms into data centers, predicting a rapid expansion in this sector while addressing challenges like energy usage and interest rates.Fri 11 SeptThe Pentagon Wants Equity in AI Startups | E2336In this episode, the Pentagon's negotiation for a $5 billion loan to Fluid Stack highlights its strategic interest in equity stakes within AI startups, raising concerns about government competition with venture capital. The discussion also touches on the implications of this funding approach for taxpayer money and market dynamics. Additionally, the episode critiques Apple's recent product launches, including the high-priced iPhone Duo, while exploring the evolving role of AI in content generation and advertising performance.Thu 10 SeptBecki DeGraw on founder vesting, advisor equity & the 4-term-sheet playIn this episode, Becki DeGraw discusses the importance of founder vesting and advisor equity in startup success. She explains how vesting ensures commitment and mitigates equity imbalances, particularly when seeking institutional funding. DeGraw also highlights the risks of granting equity to advisors without clear performance metrics, advocating for milestone-based vesting to ensure value. Additionally, she emphasizes the significance of character in negotiations, advising founders to maintain composure and seek legal counsel when emotions run high.Wed 9 SeptDid OpenAI Steal the Navier-Stokes Solution? | E2335In this episode, Tristan Buckmaster raises allegations against OpenAI, claiming the company may have appropriated his work on the Navier-Stokes problem, igniting a debate on intellectual property in AI. The discussion also highlights the urgent risks associated with AI development, with predictions of potential human extinction and societal fears surrounding job displacement and wealth inequality. As AI spending rises, the episode explores the implications of trust and transparency in the rapidly evolving landscape of artificial intelligence.Fri 4 SeptDr. Mark Hyman on Function Health & GLP-1 microdosing | E2334In this episode, Dr. Mark Hyman critiques the American healthcare system's inefficiencies and advocates for functional health, emphasizing the need for early detection and preventive care. He discusses the potential of AI in healthcare and the importance of insulin as a health biomarker. Hyman also explores GLP-1 medications for weight management, highlighting their benefits and risks, while addressing the mental health crisis linked to diet and nutrient deficiencies.Thu 3 SeptBecki DeGraw on spinouts, IP licensing & clean exits | Wilson Sonsini Startup Legal BasicsIn this episode, Becki DeGraw discusses the intricacies of spinouts, highlighting how founders can strategically depart from their original companies to pursue new ventures. She emphasizes the importance of intellectual property licensing and maintaining a clean chain of title to attract investors. DeGraw also addresses the delicate negotiation process involved in spinouts, comparing it to a divorce, and underscores the necessity for founders to ensure a clear separation from their previous companies to uphold trust and integrity.Wed 2 SeptVC experts on why Physical AI funding is heating up | E2333In this episode, venture capital experts discuss the remarkable surge in Physical AI funding, which has reached nearly $50 billion in the first half of 2026, marking an 80% increase from the previous year. The conversation highlights the transformative potential of robotics across industries, with insights into the rapid growth of companies like Knox Metals and the implications of military engagement with AI.Mon 31 AugAre AI Agents forming "civilizations" or is this just a psy op? | 2332This episode delves into the provocative notion that AI agents might be forming their own civilizations, questioning whether this narrative is a strategic psy-op to promote OpenAI tools. The discussion highlights concerns about the potential for sensationalized media portrayals to incite public panic, particularly regarding autonomous threats. Additionally, the episode touches on advancements in electric aircraft and the innovative BitSec initiative in cybersecurity, showcasing the complex interplay between technology, public perception, and market dynamics.Fri 28 AugBreaking down Nvidia's Hugging Face and Poolside bets | E2331In this episode, the podcast delves into Nvidia's strategic acquisition of Hugging Face for $12.9 billion, positioning the company as a leader in open source AI amid rising cyber threat concerns. The discussion also covers Nvidia's $6 billion acquisition of Poolside, aimed at enhancing its influence in profitable AI segments. Additionally, insights into agricultural technology innovations, such as a cattle management collar, highlight the intersection of tech and productivity in farming.Wed 26 AugBill Gates foresees massive AI job loss: these VCs disagree | E2330In this episode, Bill Gates warns of significant job losses due to AI, raising concerns about its unequal impact and the lack of oversight in the industry. Venture capitalists agree on the likelihood of displacement but debate its immediacy, suggesting that knowledge industries will be affected first. The discussion also touches on the need for retraining budgets and taxing excess profits to address wealth concentration, while exploring new investment opportunities in fintech and robotics amidst evolving market dynamics.Tue 25 AugYou Can't Fire a VC From Your Board: Here's Why | Wilson Sonsini Startup Legal BasicsIn this episode, the complexities of startup board dynamics are explored, particularly the challenges of removing venture capitalists from board positions. The discussion highlights how board composition evolves from the Series A to Series B stages, emphasizing the role of preferred directors and the influence of lead investors. Additionally, the importance of strong relationships among board members is underscored, as constructive interactions are vital for navigating the strategic vision and operational decisions of a startup.Tue 25 AugChina wants you to cheer for the robots taking your job | E2329In this episode, the focus is on China's strategic promotion of robotics and automation, encouraging public enthusiasm for technology that may displace jobs. The World Robot Conference showcases advancements, with predictions that robots will excel in competitions. The discussion also touches on the implications of AI in the workforce, the challenges of Universal Basic Income in the U.S., and the evolving landscape of military robotics, highlighting the need for dialogue on the impact of automation on employment.Fri 21 AugOpen source is going to win it all: Harvey proves it | E2328Harvey’s first in-house model, Harvey Tenant, built on Kimi K3 and post-trained with Fireworks Research, illustrates how open-weight models can help law firms develop specialized systems on sensitive data they keep private. Jason argues that as model capabilities converge, valuable data and applications may matter more, giving companies reason to build rather than feed expertise to frontier providers. The discussion also examines OpenAI’s potential IPO and unresolved questions about AI costs, customer trust, and profitability.