Trump vs Powell, Solving the Debt Crisis, The $10T AGI Prize, GENIUS Act Becomes Law

Gavin Baker argues the market’s modest reaction to firing rumors is misleading; an actual dismissal would likely trigger a far larger selloff and damage the perceived independence of the Fed (critical for inflation anchoring and global trust). He cautions that Trump may misread a –1% trial balloon as a green light, when the true risk is a broader loss of confidence in U.S. monetary governance.

Friedberg explains that with $36T in debt at an average 3.3% rate, interest costs are already ~$1.2T/year. At 5% (today’s 30-year yield), that nears $2T/year and will grow as deficits add more debt. Baker notes that for decades rising debt didn’t “matter” because rates kept falling; now, with structurally higher rates, interest will soon eclipse Medicare, Social Security, or defense—forcing action on spending, taxes, and growth.

Baker outlines a three-part path: (1) slow government spending growth, (2) modestly increase revenues (including quasi-consumption taxes via tariffs), and (3) grow faster through deregulation and productivity gains. Closing deficits should push long rates down, which in turn reduces interest costs, reinforcing fiscal improvement. But both parties have shown little appetite for real spending restraint, suggesting either a centrist third force or significant internal reform is needed.

Jason sketches a rough model: 1 billion knowledge workers in developed markets each paying ~$1,200/year for AI tools equals ~$1T in annual revenue—supporting $10T+ in aggregate market cap, before any ASI breakthroughs. Baker broadly agrees this scale is plausible, especially as AI demonstrably doubles programmer productivity and accelerates software, customer support, and sales, though he warns diffusion is slower in large enterprises than in startups.

Baker defines AGI as systems that can take economically useful actions across many domains better than an average human (drafting contracts, doing diagnoses, booking travel). ASI, by contrast, is intelligence beyond any human, leveraging all human knowledge to potentially solve problems like cancer or fusion. Friedberg reframes both as spectrum points on “leverage against complexity,” where digital intelligence compresses problem-solving time from decades/centuries to years—fundamentally changing what projects humanity can pursue.