Trump: Send National Guard to SF, China Rare Earths Trade War, AI's PR Crisis

Friedberg cites city data showing crime down ~30% citywide and ~40% downtown, homicides at a 70-year low, tents largely removed, car break-ins at a 25-year low, and net police hiring for the first time in seven years. Yet Sacks argues Market Street still resembles an open-air drug market with a concentrated Honduran fentanyl network, suggesting a targeted National Guard or federal operation could rapidly clean up the area, as Newsom did before Xi’s visit.

The hosts claim San Francisco spends roughly $700–800M annually on homelessness—around $52,000 per homeless person—largely via NGOs paid per “client,” which they argue incentivizes maintaining rather than solving addiction and homelessness. Programs like the $5M/year Managed Alcohol Program (free beer for alcoholics) are highlighted as emblematic of well‑intentioned but absurd policies. They call for cutting these funding flows and mandating treatment transitions instead.

China identified rare earths, EVs, batteries, and pharma APIs as strategic sectors in the 1990s, then used subsidies, provincial balance sheets, and quasi‑state entities to undercut global prices, drive competitors like Molycorp out of business, and build a near‑monopoly in mining, processing, and magnet casting. WTO rules allowing “developing countries” to subsidize key industries, plus U.S. environmental and regulatory burdens, enabled this outcome and left the U.S. geopolitically vulnerable.

While Friedberg is wary of permanent price floors and prefers deregulation and tax incentives, Sacks and Chamath argue that in strategic materials like rare earths, pure market logic can’t overcome China’s ability to dump supply and crash prices. They support public–private structures where the U.S. acts as buyer of last resort, builds strategic reserves (analogous to oil), and offers enough price certainty that investors can finance domestic mining, processing, and magnet production despite Chinese dumping risk.

Sacks argues that U.S. elites wrongly believed enriching China would liberalize its politics (Fukuyama’s ‘End of History’ logic), but instead created a powerful illiberal competitor, confirming warnings from Mearsheimer and Huntington. He notes China’s parallel institutions—BRICS, Belt and Road, Shanghai Cooperation Organization—and describes it as a “re‑ascending” power that historically held the world’s largest GDP ~70% of years since 1500. The solution now, he says, is a top‑level Trump–Xi ‘grand bargain’ to stabilize competition while both sides reduce strategic dependencies.

Chamath highlights three canceled hyperscaler data center projects (Google in Indianapolis, Microsoft in Wisconsin, Amazon near Tucson) amid resident concerns about higher electricity prices, water use, and noise. At a national level, media doomerism (job loss, ‘Terminator’ scenarios, OpenAI erotica panic) and a proliferation of state AI bills risk a fragmented regulatory landscape and public resistance. He argues big tech must use its balance sheets to directly benefit host communities (e.g., subsidizing power, storage, or local bills) and the industry needs better, more credible spokespeople to explain AI’s real value.

Sacks notes Q2 U.S. GDP growth was 3.8%, with ~40% attributed to AI, implying AI is the difference between a “great” (~4%) and modest (~2%) growth rate. He and Friedberg argue historical precedents (agriculture-to-industry, horse-and-buggy to Model T) show that while specific roles shrink, new, better‑paid jobs emerge, often with recruitment into new sectors preceding old-job decline. They frame AI as ‘middle‑to‑middle’ (Balaji’s term) automation that handles rote middle steps while humans remain end‑to‑end supervisors, prompt designers, and validators.