- Anthony Scaramucci argues Bitcoin is following its historical four-year cycle with a double bottom near $60K, and despite institutional demand from ETFs buffering the downside, the current bear market hasn't been as deep as past cycles (only ~50% down vs. the historical 85%).
- The NYSE's selection of Avalanche to settle trades on its new 24/7 tokenized stock platform highlights Avalanche's customizable layer-one architecture as a key advantage for legacy institutions seeking proprietary blockchain solutions.
- Scaramucci criticizes the crypto industry's political alignment with Trump, warning that the resulting Democratic backlash (a "blue wave") will create regulatory headwinds, though he remains long-term bullish on Bitcoin and core layer-one assets like BNB, Solana, and Avalanche.
Podpower Radio
Open RadioTuning in…
Episode feed
Latest episodes
- Anthropic's IPO valuation of $1.2–1.4 trillion makes it one of the world's largest companies, backed by the fastest revenue ramp ever recorded at this scale (12.3x in 3.5 years).
- The company holds mid-80s gross margins and a newly positive operating margin, with roughly $215 billion in cash covering about half of its compute commitments through 2030.
- The analyst is passing on the IPO due to aggressive valuation, compute commitment risk, sovereign AI headwinds, and LLM commoditization, preferring to wait for post-IPO price declines.
- Anthropic’s record-setting $107 billion IPO made it the world’s seventh-largest public company, but its first-day trading and governance structure drew scrutiny about investor confidence and founder control.
- Google’s new Gemini 4 Neon model impressed users with frontier-level performance and a distinctly different reasoning style, positioning Google as a renewed competitor to OpenAI and Anthropic.
- New U.S. export controls are squeezing Chinese AI firms like Tencent and Alibaba, while also raising scrutiny for chipmakers such as AMD over their China exposure.
- The conversation centers on the claim that Russell Brand’s conviction was a politically motivated prosecution, with comparisons to sentencing disparities and allegations of selective enforcement in the UK justice system.
- A major focus is the trial evidence and accuser credibility, including criticism of lifetime anonymity for complainants and the argument that the charges were weak or inconsistent.
- The discussion also highlights concerns about media and government censorship, including the role of a documentary, calls to demonetize or deplatform Brand, and penalties for speech disputing the verdict.
- The conversation highlights how a series of major tech outages (Microsoft, airlines, banks) exposed the fragility and aging infrastructure of U.S. systems, raising concerns about cyber vulnerabilities and the potential for AI to find and exploit weaknesses.
- A central theme is the alleged Chinese intelligence operation involving Nvidia CEO Jensen Huang and an employee suspected of being an access agent, illustrating how espionage can leverage personal relationships to gain access to sensitive corporate and policy information.
- The discussion covers the political fallout from the midterm elections, attributing Republican losses to factors like the Iran war, inflation, and unfulfilled promises such as releasing the Epstein files, while also touching on the rise of socialist ideas among young voters.
- Retail investors are pulling billions from Bitcoin spot ETFs, while registered investment advisors and institutions are steadily adding long-term exposure, creating a clear divide between panic selling and sticky capital.
- The Anthropic IPO, hyped as the largest ever, appears to have siphoned liquidity from crypto markets as retail chased the shiny new listing, though the guest expects this rotation to be short-lived.
- The altcoin market is maturing into a winner-take-few environment where tokens need real utility—such as RWAs, stablecoins, or yield—rather than just a TGE and hype to succeed.
- David Heinemeier Hansson argues that Linux is uniquely suited to the age of AI agents because its terminal-heavy, config-file-driven, and detailed-error-message design—historically seen as weaknesses—are exactly what agents excel at, making macOS and Windows unfit for agentic computing.
- Omarchy, an agent-first Linux distribution built largely by AI agents, demonstrates how AI collapses software development costs and timelines, enabling tiny teams to compete with large incumbents and potentially reshaping the OS market duopoly.
- Hansson stresses the importance of open-weight models and competition among frontier labs to prevent a duopoly in AI, warning that concentrated control—whether commercial or government-enforced—would be catastrophic for innovation and the open internet.
- The episode debates whether Nvidia's board was right to prioritize governance over Jensen Huang's founder-led vision, weighing the value of "techno kings" against the need for accountability at a $4 trillion company.
- Anthropic's upcoming IPO is framed as a test of whether it's a long-term investment or just a great trade, with discussion of its TPU advantage, Amazon/Google stakes, and the risk of a wild unlock.
- The conversation explores how AI-driven patching and rising leverage are creating permanent volatility and outages, and why index-fund advice may no longer pencil out for younger investors facing discontinuous wealth outcomes.
- The talk presents a new approach to integer linear programming using differentiable Dedekind sums, extending classical number theory tools to find integer points inside rational polytopes.
- Dedekind sums, characterized by modular reduction and reciprocity laws, can be computed in linear time via a GCD-like algorithm, and their higher-dimensional analogues enable efficient point-finding in rational simplices.
- The method connects to diverse applications including cryptography, the Frobenius coin exchange problem, and partition theory, with the two-dimensional algorithm conjectured to be optimal.
- MIRI has shifted almost entirely from technical alignment research to communications and governance, aiming to help the world understand superintelligence risks and figure out what to do about them, including through a new DC office.
- Key public inflection points for AI risk awareness include ChatGPT, the Hugging Face incident, DeepSeek for the national security community, and Jacob Coxon's resignation, each moving different audiences from "far mode" to "near mode" on AI risk.
- Bourgon argues that labs racing in the past doesn't mean they'll self-regulate in the future, and that external regulation is needed because AI poses a fundamentally different scope of challenge than product liability or misuse concerns.









