Wednesday, November 4
AI's Power, Money, and Trust Problem
AMD (opens in a new tab) just posted its largest and most profitable quarter in company history, driven almost entirely by data center GPU revenue, while AI inference startup Fal doubled its valuation to $16 billion (opens in a new tab) on surging demand — yet the infrastructure underpinning all of it is running into a hard physical wall. Data centers are now repurposing old jet engines to generate power as grid interconnection queues stretch years into the future. American voters made their frustration concrete at the midterms, where anti-data-center ballot measures (opens in a new tab) passed with wide margins, signaling that the AI buildout is now a live political liability, not just an engineering challenge.
North Korean hackers pulled off the largest crypto theft in history — a $6.3 billion Binance heist (opens in a new tab) — using fake job interviews to compromise insiders and then leveraging an internal AI tool to execute the attack, tactics that researchers had been flagging for years (opens in a new tab) before the breach occurred. The scale of the operation exposes how AI is now a weapon in nation-state financial crime, not just a productivity tool. Ted Schlein's Ballistic Ventures (opens in a new tab) is raising $490 million for new cybersecurity funds, a bet that the enterprise security market is about to see sustained, AI-fueled demand that existing vendors are structurally unprepared to meet.
Sam Altman (opens in a new tab) is now publicly arguing that human control must ultimately cap AI progress — a striking constraint coming from the CEO of OpenAI — while Anthropic is building a cryptographic identity system (opens in a new tab) to verify which AI model you're actually talking to, addressing the basic trust problem that emerges when model impersonation becomes trivial. These moves arrive as Palantir's blowout quarter contains a quietly concerning signal: its commercial growth is decelerating even as government contracts surge, suggesting AI enterprise adoption outside the defense sector may be softer than the headline numbers imply. Canva's simultaneous announcement of layoffs (opens in a new tab) affecting 22 percent of its staff — framed as an AI-driven restructuring — makes the same point from the bottom up: the productivity gains are real, but so is the displacement.