Hub
Regulation, liability & security
Hub for regulation, liability & security: 298 ideas from 36 episodes, organised into pages.
This area covers what the hosts say about AI regulation, the labs’ safety messaging, who pays when AI goes wrong, and security. It is mostly Sam’s thesis, relayed by Dave and Jess, and it records host opinion only, not a position of the wiki.
The most repeated ideas: Sam argues the labs’ scary narrative doubles as fundraising and policy leverage (Sam · Anthropic Wants to Slow Down. (Sep 2026) ▶ 13:57); he argues labs that cannot win on economics turn to regulatory capture (Sam · Open Source AI Wins. Now Labs (Jul 2026) ▶ 22:17); and models that find vulnerabilities make existing security fragile, a view Dave shares in part (Sam · The Real Reason Elon Wants Sp (Oct 2026) ▶ 03:12) (Dave · Amir Efrati (May 2026) ▶ 10:24). Sam also proposes that labs carry large liability cover (Sam · The Real Reason Elon Wants Sp (Oct 2026) ▶ 10:26).
The main disagreement: Brit says the safety concerns are real, not just marketing, though the US cannot slow down while China races (Brit · Anthropic Wants to Slow Down. (Sep 2026) ▶ 08:34). Dave allows that the doom media tour may show real power (Dave · iPhone Duo, Meta Muse & Why AI (Sep 2026) ▶ 32:29). Much of the lobbying material is hearsay (Dave · iPhone Duo, Meta Muse & Why AI (Sep 2026) ▶ 32:54).
Start with Regulation, liability and the doom narrative, then Regulatory capture and the doom narrative, Who pays when AI goes wrong and AI and security: the hosts’ threat model.
Pages in this area
| Page | Type | Episodes | Confidence |
|---|---|---|---|
| Regulatory capture and the doom narrative | synthesis | 18 | low |
| AI and security: the hosts’ threat model | synthesis | 18 | low |
| AI regulation: what the hosts argue | synthesis | 16 | medium |
| Who pays when AI goes wrong | concept | 6 | low |
Covered as sections of other pages: Workplace data capture → Agents at work.
Live debates
- Should states regulate AI? — see AI regulation: what the hosts argue
- Is a slowdown or review regime defensible? — see AI regulation: what the hosts argue
- Was OpenAI’s reported sandbox escape real or a stunt? — see AI and security: the hosts’ threat model
- Is a cybersecurity company buildable in 2026? — see AI and security: the hosts’ threat model
- Is the labs’ safety messaging cynical positioning or genuine concern? — see Regulatory capture and the doom narrative
- Is a company that is sold for its data set sold for scrap? — see Agents at work
Recurring questions
- Should AI labs be regulated or carry liability? — 8 answers
Blueprints
Sources
- (01)
iPhone Duo, Meta Muse & Why AI Finally Scared Dave Morin
32:29 into “iPhone Duo, Meta Muse & Why AI Finally Scared Dave Morin”
- (02)
Open Source AI Wins. Now Labs Are Running to Washington |Thinking Machines, Johnny Ive, Realtime API
- (03)
Normal People Don't Want AI Agents (And Silicon Valley Can't Accept It)
10:24 into “Normal People Don't Want AI Agents (And Silicon Valley Can't Accept It)”
- (04)
The Real Reason Elon Wants SpaceX and Tesla to Merge Now
3:12 into “The Real Reason Elon Wants SpaceX and Tesla to Merge Now”
- (05)
Anthropic Wants to Slow Down. Is It Because They're Losing?
8:34 into “Anthropic Wants to Slow Down. Is It Because They're Losing?”