OM DETTA AVSNITT
If your AI ends the world, who’s left to sue you?
En Hong (Generis AI) and Mathias Bock (Tokuma Labs) open Episode 10 with the reported leak of Anthropic’s S-1 and the strange business of disclosing existential risk to prospective shareholders. If the labs genuinely believe their technology could end humanity, why is beating the other lab still the priority?
The conversation moves from China’s AI ambitions to a more immediate problem inside law firms: juniors generating more work than their seniors can review, contracts that cover everything except the commercial point, and mistakes so convincingly packaged that nobody thinks to question them. Along the way: AI guardrails, artificial suffering, and whether a software company should get to write your moral code.
Topics Covered:
• The Apocalypse Disclaimer: Explosive revenue projections, the limits of exponential growth, and the absurdity of protecting yourself against shareholder lawsuits after the shareholders have ceased to exist.
• Safety First. Winning Firster: The tension between believing AI is dangerous and insisting your company must build it first. En and Mathias revisit the race dynamics behind the labs’ safety arguments.
• China’s Last Bus: Why apparent enthusiasm for AI might owe as much to the fear of being left behind as to optimism. Mass layoffs in the 1990s, today’s gig economy, and the historical context missing from the polling numbers.
• Distillation Versus Discovery: En challenges the argument that slowing Western AI development automatically hands China the lead. The conversation follows the comparison through Tesla, Chinese EVs, and the difference between perfecting a process and making the next breakthrough.
• The 4,000-Page Contract: Could AI draft a contract that anticipates every conceivable objection? Perhaps. Would anyone sign it? Why commercial judgment means knowing which risks to allocate, which points to concede, and when to stop drafting.
• The Productivity Backlog: Juniors hand over polished AI output. Senior lawyers inherit fifty unnecessary redlines. Clients ask where the savings are. How making documents faster can leave the whole transaction moving slower.
• Guardrails, Pain Signals, and Please: From ChatGPT smothering lecture notes in disclaimers to the debate over mistreating AI. En and Mathias disagree over where sensible safeguards end and corporate paternalism begins.
• The Mistakes Nobody Sees Yet: Hallucinated court citations make headlines. Errors buried in commercial agreements may sit unnoticed until someone wants to exit an investment or terminate a deal. What happens when plausible language pushes lawyers beyond the boundaries of their own expertise?
• On Bullshit: The distinction between lying and simply not caring whether something is true—and why AI makes the latter so much easier to produce.
Mentioned in this episode:
• Article: “Boarding China’s Last Bus” by Zilan Qian — Asterisk. Mathias’s recommendation on the history behind China’s response to technological upheaval.
• Book: Apple in China: The Capture of the World’s Greatest Company by Patrick McGee. En’s recommendation on Apple’s supply chain, China’s manufacturing rise, and the knowledge transfer that connected them.
• Essay: “On Bullshit” by Harry Frankfurt. Mathias’s recommendation on the difference between deception and indifference to truth.
Also mentioned:
• Book: The House of Morgan: An American Banking Dynasty and the Rise of Modern Finance by Ron Chernow. Referenced in En’s account of using AI to revise his teaching notes.
About Distributed Dissent:
En Hong (CEO, Generis AI) and Mathias Bock (CEO, Tokuma Labs) are two lawyers-turned-founders discussing law, finance, and AI without a corporate filter. Candid arguments, practical experience, and the books and ideas shaping how they see the industry.
