Podpower Radio

Open Radio

Tuning in…

Episode feed

Latest episodes

  • Steve Jurvetson recounts how his partnership initially voted down investing in SpaceX, with one partner warning against letting "personal passions cloud our investment judgment," before he overrode them and backed Elon Musk's last-ditch Falcon 1 launch in 2008.
  • The panel debates Anthropic's massive IPO as a landmark moment—effectively "IPOing the singularity"—and questions how public-market quarterly pressures will affect a company whose mission is long-term AI safety.
  • Discussion of AI-driven cyber threats frames the current period as one of maximum infrastructure vulnerability, where attackers can scale exploits with agents while defense remains human-in-the-loop, raising the need for "defensive co-scaling."
Atlas
  • Humanoid robots are seen as the wrong form factor for the next decade, with wheeled or alternative designs better suited for most industrial, manufacturing, and home use cases.
  • Generalized robot intelligence is advancing rapidly, but practical deployments will favor simpler, task-specific embodiments like two-finger grippers over complex human-like hands.
  • There is a growing founder shift toward ambitious physical-world ventures—robotics, manufacturing, and reindustrialization—driven by AI commoditizing software and a desire to rebuild domestic skills and supply chains.
Atlas
  • The panel agreed that capital is becoming more expensive as interest rates normalize from a 40-year low, with governments likely to resort to financial repression and money printing to manage rising debt costs.
  • Brent Johnson argued that US dollar stablecoins extend dollar dominance globally by co-opting private innovation, while Peter Schiff countered that only tokenized gold offers true stability as fiat currencies depreciate.
  • Panelists recommended owning real and productive assets—gold, equities, and income-producing real estate—to hedge against currency debasement, with central bank gold accumulation cited as a signal of gold's future role.
Atlas
  • Agents are becoming the primary users of software, demanding different design priorities like efficiency, performance, and programmability over human-oriented interfaces.
  • Microsoft is evolving Office and Copilot to support multiplayer collaboration between humans and agents, moving beyond the single-player model inherited from software engineering.
  • As AI handles more execution, human work shifts toward higher-order thinking—system design, verification, goal-setting, and coordination—rather than manual coding or content generation.
Atlas
  • Bitcoin has become boring to many longtime holders, with original believers shifting to privacy-focused assets like Zcash while traders favor protocols with transparent tokenomics and real revenue.
  • The migration of traditional finance onto blockchain rails is seen as inevitable, driven by tokenized equities and 24/7 global trading access.
  • AI agents are expected to transact heavily on blockchain rails, creating massive transaction volume and making revenue-generating protocols more attractive than Bitcoin itself.
Atlas
  • Neil Howe argues that in the crisis phase of the Fourth Turning, preparation must happen now—building skills, community ties, and retreat options before the crisis fully unfolds, citing historical examples like Athens' pre-built fleet and FDR's 1940 rearmament.
  • Howe's ETF strategy prioritizes avoiding drawdowns through cash-like holdings scaled to Shiller CAPE, trend-following and tail-risk options, plus commodity futures, real-return infrastructure, and precious metals as a bond replacement.
  • He warns that US infrastructure is highly vulnerable to cyber threats like China's Volt Typhoon, which has implanted invisible malware in SCADA systems controlling the power grid, water, and gas—and that America's complacency and lack of civil defense leave it dangerously exposed.
Atlas
  • Great PMs focus on building a shared brain by filling customer, product/market, and business context gaps so teams can make decisions independently rather than relying on the PM as decision-maker.
  • AI amplifies both strengths and weaknesses, so specialization still matters—PMs should use AI to accelerate decision-making velocity and close context gaps, not to become builders or ship code.
  • Diverse PM archetypes (artist, scientist, general manager) and diverse teams—including juniors who bring AI-native ways of working—remain critical, as experience and taste become key differentiators when everyone has access to the same intelligence.
Atlas
  • Zipline has secured major partnerships with Uber Eats and Walmart to scale autonomous drone delivery to millions of daily deliveries across the US.
  • The company is scaling rapidly to meet demand from multiple strategic partners each seeking over a million deliveries per day, requiring massive expansion of manufacturing and operations.
  • Keller argues that hardware and infrastructure companies will be the most valuable over the next decade, but investors underestimate the time and complexity required to scale them.
Atlas
  • The Frontier RSI benchmark focuses specifically on whether LLMs can improve other LLMs, distinguishing true recursive self-improvement from broader AI-improving-AI tasks.
  • The benchmark uses realistic, open-ended LLM development tasks like post-training, kernel optimization, and inference, rather than toy problems or reproducing known research.
  • The team views capabilities and alignment as orthogonal, arguing that RSI techniques can accelerate alignment research rather than conflict with it.
Atlas
  • Cybersecurity has become the top threat to global infrastructure, with AI agents enabling superhuman hacking and cross-company breaches, requiring companies to invest heavily in defensive AI and code auditing.
  • Supply chain and third-party vulnerabilities, such as a single compromised pixel on a website, can cascade into major breaches, making it critical to audit all code and own every layer of the software stack.
  • Nvidia's push for open ecosystems and open-weight models is vital for defensive security and maintaining free access to intelligence, though it faces pushback from proprietary interests and geopolitical tensions.
Atlas