Trump's Cabinet, Google's Quantum Chip, Apple's Flop, TikTok, State of VC

Rabois explains that Khosla Ventures is “input-driven,” going as early as a keynote deck with no product or metrics, while Founders Fund is “output-driven,” typically investing at $500M+ valuations. He asserts his edge is assessing whether a founder alone can build an iconic company, preferring to avoid later stages where strong metric-focused investors dominate. Actionable implication: founders seeking Rabois/Khosla-style backing should emphasize vision, personal capability, and zero-to-one thinking more than current KPIs.

Chamath, Friedberg, and Rabois argue that Trump’s billionaire/operator-heavy cabinet reflects the founders’ ideal of temporary civic duty rather than career politicians. They contend real-world P&L experience helps avoid misdesigned remedies (e.g., DOJ’s proposed Google browser divestiture) and bad mega-projects. The trade-off is potential regulatory capture, but they prefer transparent conflicts plus strict lobbying/recusal rules over today’s opaque bureaucratic revolving door.

Friedberg outlines how Google built logical qubits using arrays of physical qubits where *error rates decrease* as more are combined—an essential prerequisite for scalable quantum computation. While Willow can perform a benchmark in minutes that would take classical supercomputers ~10²⁵ years, fault tolerance is still multiple orders of magnitude from being commercially useful. For practitioners, this signals a 3–10 year window to migrate to post-quantum encryption and rethink long-term cryptographic assumptions, especially in financial and government systems.

Chamath and Friedberg note that Shor’s algorithm on a sufficiently powerful quantum computer could factor large integers and break RSA-2048 and SHA-256 (Bitcoin’s core primitive) in minutes. Back-of-the-envelope estimates suggest on the order of 4,000–8,000 high-quality logical qubits might be enough, putting industry on a “2–5 year shot clock” if progress compounds. Actionable implication: security architects and protocol designers should accelerate adoption of NIST-approved post-quantum schemes and design migration paths for long-lived secrets and chains.

Chamath and Rabois both describe iOS 18 and the latest iPhone as unreliable—crashes, broken phone calls, and an unusable Photos app—despite Apple’s world-class silicon. Rabois argues Apple once relied on Steve Jobs/Jony Ive’s singular design taste; without that, and with institutional resistance to data-driven UX testing, the company now suffers the “worst of both worlds.” Leaders without a strong taste bar should institutionalize robust user testing and metrics as a backstop instead of assuming inherited design intuition will carry forward.

Rabois cites evidence that ByteDance employees in China accessed U.S. user data and tracked journalists, contradicting TikTok’s CEO’s sworn testimony. He emphasizes China’s National Intelligence Law, which compels companies to turn over data to the CCP without courts, plus concerns about covert content manipulation and asymmetric market access (U.S. apps banned in China while TikTok thrives here). He notes overwhelming bipartisan votes followed classified briefings, implying undisclosed risks—suggesting policymakers and enterprises should treat TikTok more like critical infrastructure exposure than a benign consumer app.

Rabois says AI and crypto startups with strong teams are getting funded quickly, while non-AI enterprise remains cool. Friedberg stresses that exits aren’t blocked by markets so much as investors unwilling to IPO or sell at 60–70% discounts to 2021–22 private marks. Both argue that large VC funds cannot rely on sub-$200M M&A to return capital; they ultimately need IPOs. Founders should therefore calibrate fundraising and burn to realistic public-market multiples rather than waiting indefinitely for a “return of 2021” that may never come.