
Oct 1, 2026 · 1h 11m
AI safeguards meet America’s depleted missile defenses
10/1/26: Ro Khanna On AI Safety Bill, Bill Gates Dire Warning On Recursive AI, US Nearly Out Of Missiles
The episode links two forms of national vulnerability: weak public oversight of transformative technology and a defense system strained by sustained conflict.
- 1Ro Khanna argues advanced AI needs enforceable safeguards against recursive self-improvement, malicious use, and unchecked data-center power.
- 2The hosts challenge Bill Gates’s reliance on self-regulation while debating AI consciousness, automation, taxation, and social consequences.
- 3Christopher Leonard says the United States used a decade’s worth of THAAD interceptors in less than a year, exposing fragile production capacity.
Don't miss
Christopher Leonard reveals that the United States used a decade’s worth of THAAD interceptors in less than a year, exposing how little sustained defense capacity exists.
The brief
Ryan Grim and Saagar Enjeti open with Representative Ro Khanna’s case for regulating advanced AI, including a ban on recursive self-improvement and stronger public oversight.
The AI discussion moves from existential risk to immediate dangers such as cyberattacks, bioweapons, model misuse, regulatory capture, and the prospect of taxing automated labor.
The hosts challenge Bill Gates’s preference for self-regulation, then debate AI consciousness, sandbox failures, job displacement, and whether abundance could produce a diminished social life.
Christopher Leonard explains that the United States used a decade’s worth of THAAD interceptors in less than a year, turning a missile-defense system into a strategic liability.
Leonard traces the shortage to specialized suppliers, slow production, and changing military assumptions, raising questions about American readiness in a larger unexpected conflict.
What was said on this episode
19 statements · 6 positive · 12 negative · 1 mixed
Tech billionaires cannot be trusted to write AI safety rules.
“You cannot trust tech billionaires to write the rules to keep us safe.”
Listen at 2:54
Recursive self-improving AI should be banned.
“we need to ban recursive self-improving AI.”
Listen at 3:08
The United States should create an independent AI safety agency.
“we've got to create that AI safety agency.”
Listen at 3:53
AI policy should support more competition and open-source models.
“I'm actually for more competition and open source.”
Listen at 5:25
International agreements could require AI model safety standards and inspections.
“you could have international agreements that call for basic safety and inspection of these models”
Listen at 7:57
AI safety standards cannot be completely value-neutral about catastrophic harm.
“you can't have something that's totally value-neutral between destroying New York City and not.”
Listen at 12:33
Data centers should be banned from residential communities and farmland.
“we should be banning these data centers in residential communities and farmland and giving communities the right to say, No.”
Listen at 13:55
Medicare for All would prevent some private-insurance abuse.
“I do believe if you had Medicare for All, it would prevent some of the abuse of the private insurance companies.”
Listen at 18:41
Healthcare AI is harmful when used to deny claims, increase billing, and maximize profits.
“The problem in the healthcare system is not technology. The problem is that it's being used as a tool to deny claims, to have excessive billing, and to maximize profits.”
Listen at 19:39
Product liability is inadequate for preventing catastrophic AI harms.
“The idea that you're setting up a system whereby somebody could kill millions of people and you're gonna kind of allow that to go forward just because you can sue the person for killing the millions of people. He's right. That is preposterous.”
Listen at 26:26
AI consciousness will become a major public-discourse issue within five years.
“5 years from now, this is going to be a real no-shit part of our discourse about, you know, AI consciousness and whether these things are alive or not.”
Listen at 30:53
AI wellness concerns are already present and will shape how AIs are treated.
“I think it's here now. And whether you want to call them aliens or not, functionally speaking, that's how we're going to treat them.”
Listen at 32:07
Tax policy should shift taxation from work toward wealth.
“I think we should be taxing wealth, not work anyway.”
Listen at 36:01
A fully automated AI life will resemble passive entertainment and declining cognition.
“I think that's what a fully automated AI life will look like.”
Listen at 37:44
The United States used over nine years of THAAD accumulation in ten months.
“in 10 months of conflict in the Middle East, the United States burned through a supply of THAAD interceptor missiles that took over 9 years to accumulate.”
Listen at 43:35
The United States had 534 THAAD interceptors by 2024.
“by 2024, the United States had only 534 THAAD interceptors in total.”
Listen at 44:32
The United States used 80% of its THAAD inventory.
“We have used 80% of the entire inventory of THAADs we had.”
Listen at 53:44
The United States lacks sufficient munitions and industrial capacity for a sudden major war.
“we do not have either the stockpile of munitions to engage in a war like that quickly, and we do not have the defense industrial base that we had in the 1940s”
Listen at 1:02:23
The United States lacks enough missiles to defend its bases and regional allies.
“The reality is, at this point, we do not have the stockpile to defend the bases or the allies.”
Listen at 1:08:15
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions

Kokeland: The Lords of Easy Money
A related book-length examination of the economic forces behind automation, capital, and American policy debates.

Amusing Ourselves to Death: Public Discourse in the Age of Show Business
The book provides a useful lens for the episode’s concern that automated abundance could become passive entertainment and endless scrolling.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.