Socialists Sweep NYC, China Catches Up in Coding, AI Memory Crunch, Micron's Blowout Quarter

Summary

The episode covers the Democratic Socialists of America (DSA) sweeping New York City congressional primaries, followed by a deep debate on the political implications. The conversation then shifts to technology and markets: China’s open-source AI model GLM 5.2 is catching up to US frontier models, Micron’s blowout earnings highlight an AI-driven memory bottleneck, the economics of orbital compute via SpaceX are analyzed, and IPO updates on Anthropic, SpaceX, and Cerebras are discussed with valuation and ramp theses.

• DSA candidates win three NYC primaries, signaling a growing socialist takeover of the Democratic Party.
• Debate on social media bans for under-16s and their impact on political radicalization.
• China’s Z.AI releases GLM 5.2, an open-weight model competitive with GPT-5.5 and Opus 4.8, raising distillation concerns.
• David Sacks argues US regulatory caution is slowing AI deployment while China accelerates.
• Micron reports massive earnings, entire 2026 HBM supply sold out, DRAM bottleneck seen as most critical AI constraint.
• Gavin Baker posits that DRAM will be 30–40% of hyperscaler capex, with Micron/SK Hynix/Samsung cheap vs AI peers.
• SpaceX orbital compute economics detailed: $5B reusable Starship launch cost may undercut terrestrial data center expenses.
• Cerebras declined after breaking IPO price; Baker sees a delayed revenue ramp from a large OpenAI deal in H2 2026.
• Anthropic projected to be worth $3 trillion at public listing with >$100B revenue and 85% inference margins.