Big Beautiful Bill, Elon/Trump, Dollar Down Big, Harvard's Money Problems, Figma IPO

The Senate version removed the proposed 10-year federal moratorium on state AI regulation, and Ted Cruz’s five-year compromise also failed. Friedberg and Chamath warned that more than 1,000 proposed state AI bills could create compliance costs that smaller startups cannot absorb, potentially strengthening incumbents.

Chamath disclosed a one-gigawatt data center outside Phoenix with a $25 billion investment cycle and an expected $2–3 billion equity commitment. He said nearby nuclear power made the deal underwritable, while a new natural-gas turbine could not arrive before 2030 because of supply-chain constraints rather than technology.

Friedberg argued that DOGE-style efficiencies cannot permanently change federal spending without Congress changing appropriations law. He described legislators’ incentive to keep money flowing to their districts, while the White House’s fiscal case depends on later appropriations cuts, impoundment, tariff revenue and stronger economic growth.

The episode reported that the dollar fell more than 10% in the first half of 2025, its weakest first-half start in over 50 years. Chamath argued that continued demand for US equities, real estate and other hard assets can outweigh currency erosion, making American innovation—not the dollar alone—the longer-term constraint.

Harvard borrowed $1.2 billion amid federal-funding uncertainty after the administration canceled more than $2 billion in research grants, according to the episode. Chamath argued that its private-equity allocation, roughly doubled from 20% to 40%, could become a liquidity problem if an estimated $1 billion annual funding gap persists.

AI may weaken universities’ monopoly on instruction, but the harder problem is replacing the credential as an employer’s talent filter. The hosts proposed projects, internships, internal training and AI interviews, while Chamath warned that late bloomers could lose opportunities when institutional brands no longer provide them an initial doorway.

Grammarly acquired Superhuman, which had raised $114 million, reached an $825 million peak valuation and reportedly generates about $35 million in annual revenue. Together with Grammarly’s earlier Coda acquisition, the transaction points toward broader AI workplace suites consolidating functions that previously lived in separate productivity products.

The deeper software risk discussed was that AI could eliminate the specialist jobs that create demand for entire categories of SaaS, not merely compete with individual products. Figma remains strong—$95 million Q1 free cash flow and roughly 130% net retention—but its long-term valuation depends partly on how much foundational models eventually absorb.