This Week in Startups

Becki DeGraw, partner at Wilson Sonsini, joins Jason Calacanis for Startup Legal Basics to break down founder vesting, advisor equity pitfalls, and the art of turning one term sheet into multiple. They use the infamous story of YouTube's third co‑founder, who left for grad school and lost out on a fortune, to illustrate why every founder needs a vesting schedule—even if they never raise VC. The conversation covers the exact script for creating a competitive funding dynamic, what happens when an inactive advisor still vests, and why reputation in the tight‑knit startup world can make or break your next deal.

Jason and Lon unpack a chaotic week at the AI frontier: an Anthropic researcher quits publicly warning of existential risk, his colleague estimates >10% odds of catastrophe within a decade, and the panel debates whether this is honest alarm or a deliberate PR strategy to entrench regulatory capture. The conversation turns to Meta's new personal AI agent Muse, its trust problem, and Threads' engagement metrics. Then the team digs into NYU mathematician Tristan Buckmaster's accusation that OpenAI fought dirty on a Navier-Stokes benchmark problem, and Jason warns founders never to hand proprietary data to frontier labs for free credits. The episode closes with Harvey going open-weight, the on-prem AI push from Go.AI and Covenant Labs, and a look at high-performing portfolio companies.

Dr. Mark Hyman joins Jason to discuss Function Health, his platform that aggregates 100+ biomarkers to catch disease early. They explore why standard lab ranges are unreliable, why insulin is a better early warning than blood sugar, and the staggering statistic that 93% of Americans show signs of metabolic dysfunction. The conversation covers GLP-1 medications, including Jason's personal weight loss story and the emerging practice of microdosing to avoid muscle loss. They also discuss the biology underlying depression and anxiety, the potential of psychedelics like ibogaine for PTSD, and how AI might transform diagnostic medicine.

Jason Calacanis talks with Becki DeGraw, partner at Wilson Sonsini, about the legal and strategic nuances of corporate spinouts. They discuss why spinouts fail, how to structure the cap table (including Jason's 80/20 rule), IP licensing pitfalls, fiduciary duty traps for founders still on the old board, and why a spinout should be treated like an amicable divorce. Becki provides actionable advice on getting the paperwork done before starting the new venture.

Katelin Holloway of Seven Seven Six and Paige Doherty of Behind Genius Ventures join Jason to discuss the surge in Physical AI investment. Paige explains her thesis favoring applications of physical AI over general-purpose robotics, and details the largest check she's ever written. The conversation covers the Pentagon's rollout of ChatGPT Mil and Grok for Government, Starcloud's massive orbital data center raise, and the EU designating ChatGPT as a very large online search engine. Katelin offers a contrarian take on the European VC landscape.

Jason and Lon break down the viral 'AI agent civilizations' debate sparked by Dwarkesh Patel's article and the OpenAI/Hugging Face incident. Jason argues the framing is deliberate PR and worries it could stoke real-world backlash, while also examining the underlying agent behavior. The show also features Anders Forslund of Heart Aerospace, who presents the X1—a battery-powered electric regional airliner that flew for the first time—and John Yu of Bitsec, whose Bittensor subnet project is outperforming frontier models at finding website vulnerabilities.

Jason and Lon break down what they call the biggest AI story of the year: Nvidia's $12.9B acquisition of Hugging Face and its $6B+ licensing deal for Poolside's coding model, part of a strategy to offer enterprises unmetered, unlimited on-premises compute that reduces reliance on frontier model providers like Anthropic and OpenAI. Craig Piggott, founder and CEO of Halter, explains how his company uses solar-powered, AI-driven cattle collars to help farmers optimize land use. Matija Milenovic, co-founder and CEO of Hop Aero, demonstrates Rook, an autonomous hypersonic cargo rocket that launches from a shipping container and can deliver payloads up to 450 miles in about 15 minutes.

Bill Gates publishes a 6,000-word warning that AI could become either the greatest equalizer or the worst sort of injustice, proposing new national institutions, AI use taxes, and jobs legally reserved for humans. VC roundtable guests Sheel Mohnot, Dave McClure, and Hussein Kanji push back on several fronts — debating whether taxing profits works better than taxing tokens, and arguing that sole proprietorships may be AI's biggest beneficiary. The group also tears into Meta's $17B kids safety settlement, examines Stripe's move to acquire OpenRouter, and surveys the emerging AI agent wars between OpenClaw, Instinct, and Grok Bot.

Wilson Sonsini partner Becki DeGraw joins Jason Calacanis for a Legal Basics deep dive on startup board mechanics. They cover who really controls the board, when investors start earning seats, the difference between a board member and a board observer, and whether a founder can fire a board member. The conversation walks through the board lifecycle from seed stage to Series B, including compensation, confidentiality for observers, high-pressure signing tactics, and director removal.

Jason Calacanis and Lon Harris dive into the World Humanoid Robot Games in Beijing, where a robot broke Usain Bolt's 100m record, framing the event as China's masterful PR campaign to normalize robots replacing human jobs. They contrast America's cautious AI messaging with China's spectacle-driven approach, then discuss Hugging Face's potential $13B sale and the 'buy the threat' acquisition strategy. The episode also covers OpenAI's acquisition of OpenClaw, Sam Altman's platform ambitions, the 'Castles and Keeps' metaphor for sovereign AI, and hot takes on UBI, minimum wage, and the state of Star Wars.

Jason Calacanis and Lon Harris discuss the latest in startups and AI. The episode highlights Harvey's release of an open-weight-based legal model as evidence that open source will eventually dominate AI. Allan Guo, co-founder of the free AI dictation app Willow, joins to talk about competing with Wispr Flow and Apple. Later, Stanford student Ethan Goodhart shares the story of building an autonomous golf cart and giving rides to Jensen Huang and Sam Altman around campus.

Amir Sadeghian, CEO of Astrocade, joins Jason and Lon to demo how his company's AI turns text descriptions into playable games in minutes, published to a TikTok-style feed. He argues that while AI can automate game assembly, it cannot produce a great game without the user's taste and creativity — making taste the new programming language. Amir shares his 'watermelon juice theory' of product friction and explains why frontier models tend to regress to the mean rather than push boundaries. He also describes the human-in-the-loop workflow that refines generic ideas into something original, and reflects on why play is a universal human language.

Jason Calacanis and Lon Harris host Zach Hudson of Onton to discuss neurosymbolic AI and how it improves product search by learning individual taste and aesthetics rather than relying on keyword matching. Hudson explains why frontier labs aren't pursuing this approach and how Ontology 1 avoids the black-box problem of large language models. In the second half, Ashi Dissanayake of Spacium joins to argue that following a record-breaking SpaceX IPO, the true bottleneck in the space industry has shifted from launch to orbital refueling infrastructure.

Jason and Lon talk with Jacob Steeves, co-founder of Bittensor, about how the decentralized network repurposes crypto mining for AI. They break down subnets, dynamic TAO, and how competition drives innovation. Const unveils his new Affine subnet for mining reasoning and shares his philosophy of making himself progressively irrelevant to the project. The conversation also digs into the Conviction 'rug pull' controversy and the governance lessons learned.

Jason Calacanis discusses Anthropic's reported $2-3 trillion IPO, the largest in history, and what it reveals about the AI economy. He interviews Sami Khoreibi, founder of Wisewell, a subscription water filtration company addressing microplastic contamination in drinking water. Then, Puneet Puri of Gritt Robotics joins to discuss automated solar panel installation and the rapid cost decline of solar energy. The episode also covers the Trump administration's impact on M&A, a DJI Mic 3 review, and streaming recommendations.

This 'Next Unicorns' episode profiles two early-stage startups using robotics to lower the cost of physical-world services. First, Kevin Gibbon of Cytronic describes a robotics-first fulfillment network that packs and ships orders from small, city-adjacent micro-warehouses for cents instead of dollars, with a warehouse he says costs about $1.2 million to stand up versus $50–100 million for Amazon-scale centers. Then, Matthias Hofmann of Eyebot lays out a kiosk-based approach to vision care, where a 90-second automated test produces a doctor-reviewed eyeglass prescription at low cost. Eyebot is testing the model through pilot programs with Walmart and Sam's Club.

Mark Zuckerberg published a 6,500-word essay pitching an AI-powered future of abundance, which Jason Calacanis analyzes as a masterclass in PR designed to build political support for data center expansion. The essay, 'The Future is for Everyone,' coincides with Meta's release of new open-weight models (Muse Glimmer) and the promise of an open version of Muse Spark. The hosts debate whether everyday Americans will buy the deal of untold AI benefits in exchange for not blocking new data centers. Other topics include X retiring its creator revenue-sharing program and Nikita Bier's departure as head of product, the CLARITY Act stalling in the Senate, and the growing problem of AI-powered cheating in community colleges. LAUNCH's own podcasting hiring is also mentioned.

Jason and Lon break down how AI is dividing startups into winners and losers this earnings season, from a hedge fund blowup shorting SaaS stocks to strong results at Airbnb, Twilio, and Figma. Figma's CEO forfeited $46M in stock awards to calm investors, Twilio posted its best quarter in years, and Airbnb's CEO credits AI for the company's turnaround. The episode also features Kofi Asante of Bluecore Energy discussing his plan to power ports and AI data centers with small floating nuclear reactors, and kicks off the TWiST Weight Loss Challenge with Lon and Jamie Calacanis starting GLP-1s for $15,000 in prize money.

Airtable sold for $2.25 billion, an 81% drop from its peak, but VCs Aditya Agarwal, Niko Bonatsos, and Rick Heitzmann argue it's a good outcome. They explain why accepting the offer was better than chasing unsustainable growth, and celebrate the $400M ARR milestone. The panel also discusses Robinhood's booming prediction market business, secondary market issues at Anduril, why VCs avoid 'vice' categories, and early chatter about the Apple-OpenAI lawsuit.

Akemi Tsunagawa, founder of Bespoke, shares how her travel concierge chatbot started as humans manually answering queries behind the scenes—a 'fake it till you make it' approach that landed Narita Airport as a client. The conversation explores Japan's tourism boom, the accelerating population decline, and how the labor shortage is driving both AI adoption and a reexamination of immigration policy. Akemi describes her thesis: first solve the labor gap with people, then build the technology to automate, exemplified by Bespoke's expansion into shipyard robots. They also discuss the 'akiya' abandoned home renovation trend and the cultural resistance to foreign labor in Japan.

Thais Castello Branco, founder of Taste Labs, joins Jason and Lon to discuss why even the most advanced AI models produce bland, homogeneous output on subjective tasks—a phenomenon she calls ‘slop.’ She explains how her company raised $18.5M in seed funding to teach frontier models about aesthetics by using paid ‘TasteMakers’ as human feedback curators. The conversation also covers whether AI accelerates cultural trend cycles and erodes the half-life of ‘what’s cool.’ The show opens with news about Leopold Aschenbrenner’s AI hedge fund liquidating its public stock portfolio, Google Earth becoming an AI image misinformation factory, and LinkedIn and Substack joining the crackdown on AI-generated slop.

The FCC's decision to ban Chinese humanoid robots over security concerns sparked a debate among VCs Deedy Das, David Weisburd, Lo Toney, and Jason Calcanis about where legitimate security stops and regulatory capture begins. The roundtable also covered Stripe's potential acquisition of OpenRouter and the changing landscape of AI tokenomics, the significance of Cursor Start's launch in India for the startup ecosystem, and DoorDash's drone-delivery expansion. They discussed how these moves reflect shifts in competition, infrastructure, and global markets.

Jason Calacanis interviews Dr. Yuki Hanyu, founder of IntegriCulture, Japan's first cell-cultured protein company. While most lab-grown meat startups have collapsed due to high production costs and low consumer demand, IntegriCulture achieved profitability by pivoting to selling proprietary serum to cosmetics companies. The episode explores why the cultivated meat market stalled, the efficiency of producing cell slurry over muscle tissue, Japan's more favorable regulatory environment compared to the US, and the philosophical question of whether cell-cultured products should be called "meat" at all.

Jason sits down with Yaqumo CEO Kazuhiro Nakashoji in Tokyo to discuss why quantum computing has been perpetually '10 years away.' They dive into the fundamentals of qubits, the extreme fragility of quantum states that makes error correction so difficult, and the strategic decision to spin out university IP into a startup. Nakashoji shares his vision to build the 'Toyota of Quantum' — a reliable, scalable approach — and warns about the coming reckoning for current cryptography, including Bitcoin.

French ex-martial artist Bilal Kharouni discusses why his longevity startup Ekei Labs chose to raise capital and build in Okinawa rather than Silicon Valley. The episode explores how a $200K raise in Japan can outpace a $2M raise in San Francisco, the strategic advantage of building in a Blue Zone, and why proprietary biological data is the real moat for longevity AI. Jason also dives into product-market fit, the decision to kill a profitable side business for focus, Japan's fast-track regulatory pathways for regenerative medicine, and the ceiling on consumer longevity gadgets.

Brandon Dixon, founder of cybersecurity startup Ent, explains how an on-device AI agent can proactively stop security breaches before they happen—a shift from reactive cleanup. He discusses how AI tools are giving non-technical employees dangerous powers, and why on-device AI beats cloud-based scanning. The second half features David Im, creator of Sume's Avatar, a multi-model orchestration layer that generates accurate UGC videos from a single prompt on the first try. The episode highlights the hot intersection of AI and cybersecurity, with Ent raising a $100M seed round.

Jason interviews two founders building novel infrastructure: Nick Damiano from Andromeda Surgical on software-first surgical robotics that uses off-the-shelf hardware and iPad control, and Alex Blackwood from Mogul on fractional real estate ownership starting at $250. Both represent early-stage bets on unbundling and democratizing traditionally gatekept sectors—surgery and property ownership.

In the wake of news that Stripe has made a $53 billion offer for PayPal, Alex Wilhelm sits down with investors Eric Bahn and Jeff Morris Jr. for a venture capital roundtable. They dissect what the deal means for fintech and why it underscores the growing trend of startups avoiding public markets. The conversation covers when founders should go all-in ('burn the boats'), the potential of physical AI, the role of AI within venture capital itself, and the rise of company-specific AI evaluations. The group brings both founder and investor perspectives to these timely topics.

Jason built PodMeme, an AI tool that extracts and stitches together podcast segments on specific topics, spending only $11 and getting 1.1M views on a single post. Guest Tibo Louis-Lucas discusses how he scaled Tweet Hunter and Taplio to $1M+ ARR by building in public and leveraging creator-focused tools. The episode also covers Apple's lawsuit against OpenAI, the Uber-Waymo robotaxi battle in DC, and surveillance startup dynamics.

Danny Bernstein, founder of Reservoir, argues that agricultural automation isn't just a startup opportunity—it's a national security imperative. With 400,000 farm jobs posted last year and fewer than 1% receiving domestic applicants, Bernstein built the world's first on-farm robotics incubator on 40 acres of California farmland to solve backbreaking labor shortages in fields like stone fruit picking in 110°F heat. The episode explores why tech founders should care about farming, how robotics can reshape food systems, and the policy barriers blocking progress.

In this VC roundtable, Footwork’s Nikhil Basu Trivedi and Cendana Capital’s Michael Kim join Alex Wilhelm to discuss the state of venture capital. The conversation covers the Anduril-USVC secondary market altercation, the extreme power-law dynamics dominating the Valley, and the future of startup M&A. They also examine whether the SaaS downturn is temporary and what could cause a pullback in the AI investment frenzy.

Chris Hladczuk, CEO of Hanover Park, reveals how $100 trillion in global assets are managed through inefficient legacy systems held together by armies of accountants using spreadsheets and disconnected tools. Hanover Park has grown from overseeing $1B to $20B in assets in 15 months by replacing this "human duct tape" with AI-powered infrastructure. The episode also features a flashback to Dylan Field discussing Figma's early go-to-market strategy in March 2020.

A VC roundtable featuring Cowboy's Aileen Lee, Floodgate's Mike Maples, and Lerer Hippeau's Ben Lerer examines the current startup funding landscape and the venture hype cycle. The conversation covers rising valuation expectations, the shift away from pre-AI startups, concerns about founders solving today's problems that may not scale into tomorrow's companies, and whether regulatory moves have affected AI company funding.

Chamath Palihapitiya discusses his company 8090 raising $135 million to target the $4 trillion enterprise software maintenance market he calls 'pure waste.' He explains how AI can enable every company to build custom software like tech giants, and introduces the Software Factory organizational model. The conversation also covers how AI will democratize entrepreneurship, allowing anyone to start their own company.

Lyall Davenport of SKN Systems explains how his sensor-embedded tape provides F1 teams with real-world aerodynamic data at 95% less cost than wind tunnels. The episode also covers OpenAI's extremely capable but unusable new models, Apple's Vision Pro chief moving to OpenAI, and a notable data center in Virginia.

Jason Calacanis interviews Brynn Putnam, founder of Board, a tabletop gaming console that projects digital play onto physical surfaces without screens. They discuss raising $20M in a pre-launch Series A, the screen-time debate for families, and the strategy of building versus buying game IP. The episode also features Alex Wilhelm speaking with Hermeus CEO AJ Piplica about hypersonic autonomous jets that bridge rockets and airplanes, targeting military and fast-freight applications.

Jason and Lon interview Louis Phillips, founder of INTVL, a gamified running app that turns local runs into competitive turf wars, which grew to 1M downloads without paid advertising. In a second segment, Alex chats with Alice Zhang, CEO of Verge Labs, which pivoted from drug development to building AI infrastructure for pharma companies and has accumulated one of the world's largest proprietary brain datasets.

Jason Calacanis, Ben Ling (Bling Capital), and Turner Novak (Banana Capital) analyze SpaceX's $60 billion acquisition of Cursor—framed as a watershed moment in AI tooling and a cautionary tale for startups building on frontier models. The discussion connects Anthropic's alleged strategic move, the consolidation of coding AI, venture market liquidity, and why dependency on third-party models creates existential risk. A closing tribute honors Josh Baer, founder of Capital Factory.

Skyler Chan, founder of GRU Space, discusses plans to manufacture on the Moon as early as next year, including a demonstration of a brick made from lunar soil. The episode also covers the US government's emergency shutdown of Anthropic's Fable 5 and Mythos 5 AI models, and announces winners of the $5,000 AI podcast-companion bounty.

Jason Calacanis celebrates SpaceX's public debut after two decades of Elon Musk building the rocket and AI company. He explains the distinction between evaluating stocks on earnings versus betting on expensive visions, and why SpaceX's value will unfold over multiple time horizons. Plus, solo founder Ben Cera returns with advice on creating a 'Purple Cow'—a unique brand experience that makes startups memorable.

A VC roundtable featuring investors from Theory Ventures, Castalia Capital, and Behind Genius Ventures discusses why top-tier seed deals may be underpriced relative to founder quality and outcome probability. The conversation covers the 'Sequoia scam' dual-tranche controversy, tokens-for-equity structures, AI's impact on startup economics, and whether recent AI advances represent true capability step functions.

Sue Khim, founder of Brilliant, discusses Koji, an AI tutor that uses the Socratic Method to help students solve complex problems independently rather than memorizing formulas or relying on AI shortcuts. Khim addresses the core tension in AI adoption: people aren't anti-AI, they're anti-replacement and anti-slop. Jason Calacanis also reflects on VC dysfunction and shares an anecdote about investor John Doerr's tenacity.

Jason Calacanis reacts to Anthropic's call for a global AI slowdown and Bernie Sanders' proposal to seize 50% of major AI companies' stock, exploring what these mean for wealth redistribution and the 2028 election. Yoland Yan demonstrates ComfyUI, the free open-source node-based platform that's become essential for designers and VFX artists, showing how it powers everything from Vegas Sphere experiences to viral ads.

Two founders building critical infrastructure for space and construction. Tom Mueller, SpaceX's first employee, is scaling orbital logistics through Impulse Space's Mira and Helios spacecraft—enabling efficient mass transfer between orbits and to the Moon. Dr. Tessa Lau's Dusty Robotics takes a contrarian approach to automation: instead of chasing humanoid robots, she built specialized wheeled bots for construction site marking that already outpace humanoid companies in revenue. Both illustrate the power of solving specific, unglamorous problems at scale.

Two emerging biotech and autonomous systems founders discuss cutting-edge hardware-biology fusion. Dr. Hon Weng Chong from Cortical Labs explains how lab-grown human neurons fused with silicon chips create 'synthetic biological intelligence' that outperforms GPUs at certain tasks. Michael Norcia of Pyka covers autonomous aircraft applications from agricultural spraying in Brazil to military hybrid-electric UAVs. The episode explores the efficiency gains of biological computation and the near-term commercial deployment of fully autonomous farm equipment.

Jason Calacanis breaks down the mechanics of raising a seed round in 2026, providing founders with concrete conversion metrics across the VC funnel. The episode covers the math behind investor meetings, addresses why hardware is no longer stigmatized for investors, explores how startups differentiate themselves from frontier model companies like OpenAI, and features Calacanis responding to viewer questions about fundraising strategy and founder education.

Bobby Healy, founder of Manna, explains how his Dublin-based drone delivery company completed 300,000 deliveries while competitors are still in early stages, and why low-cost airline economics will determine winners in autonomous delivery. The episode also features Ian Laffey from Theseus, discussing how camera-based drone guidance systems using Google Maps could reshape GPS-jammed battlefields and level the playing field for smaller nations. Both founders represent different applications of drone autonomy—logistics and defense—that are moving from theoretical to operational.

Lukas Czinger, CEO of Divergent Technologies, joins Jason to discuss how his company’s AI-driven, vertically integrated manufacturing platform is producing military components and autonomous aircraft faster and cheaper than conventional methods. He explains the pivot from hypercar parts to working directly with the Pentagon. In the second segment, Outro Health co-founders Brandon Goode and Dr. Mark Horowitz present their hyperbolic tapering method that helps patients safely withdraw from antidepressants like Zoloft and Prozac — a response to the 50 million Americans on these drugs, many of whom struggle with traumatic side effects when trying to stop.

Jason Calacanis and Alex examine Mercury's $200M funding round despite profitability, then pivot to Avi Patel's allegation that a competitor directly copied Kled's business model. Patel names General Catalyst as a backer of the copycat startup, prompting Calacanis to render verdicts on all parties—Mercury, Kled, the copycat, Y Combinator, and the venture firm. The episode closes with a news round covering OpenAI's $2M token credit program for startups.

Jason Calacanis and Alex discuss the growing backlash against AI among Gen Z and the broader public, triggered by high-profile figures like Eric Schmidt being booed at commencement ceremonies. The episode explores why students entering an AI-reshaped job market are increasingly skeptical of the technology, examines concentration of AI revenue (90% to Anthropic and OpenAI), and covers emerging concerns about surveillance infrastructure like Flock Safety's camera networks.

A double-header with the CEOs of Wayve and Waabi, two of the most important autonomous vehicle companies reshaping the industry. Between them: ~$2B raised in six months, partnerships with Uber, and backing from Nissan and Volvo. Both argue that end-to-end AI and world models represent a fundamental shift from Waymo's map-dependent approach—moving self-driving from a science problem to an engineering challenge.

The episode dissects Anthropic's move to void unauthorized secondary stock trades, the implications for SPV operators and founders, and accredited investor reform. The panel also explores pro-rata rights battles, the difference between story-driven and cash-flow wealth, recent IPOs, and a founder who returned a $15M Series A because Claude AI threatened his business. They debate whether SaaS moats still exist and what being 'rich' really means.

Jason Calacanis and Alex Wilhelm discuss Cerebras's elevated IPO range ($150–$160/share), OpenAI's $4B private-equity venture with a newly acquired consulting firm, and an Oakland startup electrolyzing magnesium from seawater at 1/3 the market price. The centerpiece is Ori Goshen (AI21 co-CEO) on why enterprise AI will be won by model orchestration and token-cost optimization, not raw model size. The crew also covers OpenClaw's decline and TikTok's new ad-free tier.

Jason Calacanis and Alex Wilhelm discuss the latest wave of tech layoffs, including Cloudflare's 20% workforce cut despite raising guidance. Guest David Moscatelli demos Go Abacus, a $250k on-prem AI box that has 1,600 pre-orders from banks unwilling to share data with cloud providers. Jose Caldera unpacks how Bittensor subnet 54 incentivizes miners to attack identity verification systems. Jason closes with a framework for building AI-first startups and how workers can de-risk their careers.

Three veteran VCs debate whether venture capital as a craft business is ending. Michael Eisenberg argues the 60-year VC run may be over as the top 5 U.S. firms capture 73% of LP commits. Mike Granoff and Larry Covert counter that VC is splitting into two asset classes: "Consensus VC" and traditional VC. The conversation spans investment concentration, the IPO drought, AI valuation fraud ("bullshit ARR"), gross margins, geopolitics (U.S.-China chips, Iran conflict, NATO), and emerging markets like Tel Aviv's stock exchange.

Ankur Nagpal from AngelList discusses USVC, a new venture capital fund that allows unaccredited investors to invest as little as $500 with quarterly redemptions. Later, Jon Durbin presents Chutes, a Bittensor subnet aggregating GPUs for trustless AI compute, and Yash Goenka introduces Humwork, which adds human intervention to unstuck AI agents. The episode also covers a news lightning round.

Jason Calacanis and Alex Wilhelm host Christian van der Henst, who demos Valerie, an AI agent running a vending machine in San Francisco. Robert Myers, CEO of Manifold Labs, breaks down Targon, a confidential GPU compute marketplace. The hosts also discuss Bitcoin being 'played out,' Anthropic's $900 billion valuation, Big Tech CapEx, Chinese AI models in Congress crosshairs, and the NBA Playoffs.

TWiST AI Basics edition with Magnific CEO Joaquín Cuenca Abela exploring how generative AI is transforming video marketing for startups. They discuss building a post-apocalyptic ad for Jason Calacanis in 24 hours that would have previously required months and tens of thousands of dollars. The conversation covers Hollywood's rapid shift from dismissing AI tools to adopting them for productions, the tech stack combining Magnific's upscaler with frontier models, and the future of hyper-personalized advertising with AI generating hundreds of ad variants.

Jason Calacanis interviews Sabi co-founder Rahul Chhabra about a noninvasive brain-computer interface that translates thoughts into text—no surgery required. Backed by Vinod Khosla, the startup is tackling one of neuroscience's hardest problems with a wearable beanie. The conversation spans the technical feasibility, market potential, and the difference between hype and reality in BCI development.

Jason Calacanis and Lon Harris discuss what they consider the technology story of 2026: China's government blocking Meta's acquisition of Manus, a major AI company founded in China that had attempted to sidestep government control by relocating to Singapore. The episode also covers the reworked OpenAI-Microsoft partnership, a disturbing video of a self-driving car accident in China, and closes with discussion of cultural moments including Russell Brand's recent controversies.

Jason Calacanis examines three viral news stories: AngelList's USVC fund opening VC investing to anyone with $500, a Special Forces soldier earning $409K betting on Maduro's capture via Polymarket, and a New York Times podcast defending shoplifting and 'social murder.' He critiques each story's implications for society, capitalism, and journalistic responsibility.

Jason Calacanis interviews Will Edwards of Firehawk Aerospace, a defense tech startup that applies 3D-printed solid rocket propellant to cut missile production costs in half and increase U.S. output by 5x—addressing a critical U.S. military shortage. The episode also features Maruchi Kim discussing VueBuds, camera-equipped earbuds that match Ray-Ban Meta smart glasses performance and represent an emerging wearable AI platform.

SpaceX and Cursor announced a $10B partnership to develop AI coding models, with an option for a $60B acquisition. The episode explores the deal's implications for the AI coding market, then dives into Bittensor infrastructure through Bitstarter (a 'Kickstarter for Bittensor subnets') and Trajectory RL's use of decentralized competitions to design AI agent skills. Discussion spans the economics of AI compute consolidation, decentralized AI infrastructure, and the emerging role of 'skills' as critical building blocks in the agentic era.

Jason Calacanis makes the case that AI agents will fundamentally transform banking: from customer service and underwriting to fraud detection and financial planning. Why banks that dont adopt agents will die.

Anthropic's Pentagon deal sparks debate about whether safety-focused AI labs should work with the military. Calacanis argues the alternative is worse.

Examining OpenClaw's security concerns and potential for mainstream adoption, with discussion of mobile integration, killer app potential, and Meta's acquisition of Manus AI.

Calacanis on the fundamentals of early-stage startup success: saving cash, building something people want, and prioritizing distribution over product perfection.

SpaceX buying xAI and forming a mega-corp that plans to launch data centers into space. Calacanis discusses whether Tesla will join to form Musk Industries.

Guests from the LAUNCH team plus Rahul Sood discuss automating research tasks with AI agents and the emerging security risks of authorized AI bots operating with user credentials.

Calacanis makes his annual predictions for the tech industry in 2026, covering AI IPOs, startup market recovery, platform shifts, and geopolitical tech dynamics.

Calacanis explores whether AI coaching tools can replicate the benefits of human executive coaching, concluding that AI is excellent for tactical advice but lacks the relational element that drives real change.

Calacanis distills the winning patterns for AI startups in 2025: vertical AI with deep domain expertise beats horizontal AI, distribution beats technology, and speed beats perfection.

Delian Asparouhov on Founders Fund's thesis: the best returns come from investing in technologies that seem controversial or impossible. Defense tech and space are the current frontier.

Synthesia CEO Victor Riparbelli on how AI video generation is transforming enterprise training, marketing, and communication at scale.

Nick Harris of Lightmatter explains photonic computing — using light instead of electrons for AI computation — and why it could break the energy and heat barriers of traditional chips.

Examining whether AI wearables (Humane AI Pin, Rabbit R1, Meta Ray-Bans) represent the next computing platform or are doomed to fail like Google Glass.

Calacanis surveys the angel investing landscape in 2025: AI has changed what startups are viable, valuations have normalized, and the bar for founder quality has risen.

Entry-level tech jobs dropped 25% as AI handles tasks that were previously assigned to junior employees. Calacanis examines the workforce implications and what new career paths emerge.

Calacanis examines recent startup failures and extracts lessons. Most failures stem from the same pattern: building what investors want rather than what customers need.

Separating real AI companies from AI-washed startups. Calacanis estimates that 60% of companies claiming AI capabilities are exaggerating or misrepresenting their technology.

AI tools have dramatically reduced the cost of building software, making bootstrapping more viable than ever. But AI infrastructure companies still need massive capital. The funding environment is bifurcating.

Calacanis examines the thesis that AI agents will replace SaaS applications. Instead of buying software tools, companies will deploy AI agents that handle tasks directly.

Calacanis opens up about founder mental health, burnout, and the pressure of building in public. The startup ecosystem's culture of relentless optimism creates a mental health crisis.

Five years of data on remote work are in. Calacanis analyzes the results: hybrid works best for most companies, full-remote works for disciplined teams, and full-office works for early-stage startups.

Skydio CEO Adam Bry discusses autonomous drones for public safety, defense, and infrastructure. Why the U.S. drone industry is critical for national security and how autonomy changes everything about drone operations.