Sep 12, 2026 · 30m · tbpn
Washington Hones in on AI Safety, What Are AI Doomers Proposing | Diet TBPN
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of Diet TBPN, John Coogan and Jordy Hays dissect concrete AI safety proposals—from the AI 2040 roadmap's hardware compute limits to aggressive legislative bans—while evaluating the practical feasibility, economic realities, and philosophical questions surrounding emerging artificial intelligence.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 97.9% of the talking time here. How this is scored →
The hosts as informed peer 5.8 Guest teaching 0.6 Guest disagreement 0.8 The hosts pushing back 1.0
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 22:25 Cosgrove pushes back on totalizing doom logic with the bunker analogy
Cosgrove directly challenges Coogan's reasoning by pointing out that people build bunkers for nuclear war, whereas AI doomers take no practical protective actions because they treat the threat as totalizing.
Hardest push from the hosts ▶ 21:57 Coogan challenges Tyler's demand for falsifiabilityCoogan rejects the premise that existential risk claims must be strictly falsifiable in financial markets, arguing that decades of Cold War nuclear policy functioned without payout-linked contracts.
Biggest teaching moment ▶ 24:19 Cosgrove explains Google's complete fly connectome mappingCosgrove informs the hosts of the exact technical breakthrough where Google mapped every neuron in a fruit fly's 3D nervous system to enable biological decision emulation in software.
The host holds their own ▶ 23:05 Coogan deconstructs Christiano's leveraged bond tradeCoogan demonstrates financial literacy by parsing Paul Christiano's short on 30-year Treasuries as equivalent to holding a conventional mortgage rather than a legitimate hedge on existential catastrophe.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Deconstructing the Concrete Proposals of AI 2040 | 7 | 0 | 0 | 0 | Coogan dominates the segment breaking down the tactical proposals in the AI 2040 plan, detailing hardware caps, Faraday cages, and 1 Mbps bandwidth throttles. The guests do not participate beyond occasional passive affirmations. | |
| Enforcement Feasibility, Compute Scale, and Secret Labs | 6 | 0 | 0 | 1 | Hays and Coogan debate enforcement feasibility and underground labs compared to nuclear non-proliferation. Coogan cites Ilya Sutskever's SSI cluster needs to illustrate that massive compute scale remains physically observable. | |
| The 2040 Superintelligence Timeline and Economic Trade-Offs | 6 | 0 | 0 | 0 | Coogan outlines the 2035 expert-level capability and 2040 superintelligence timeline, contrasting it with 2028 predictions. He critiques p(doom) debate distractions and articulates the economic trade-offs of safety regulations. | |
| Evaluating Physical Robotics Risks and Expert Perspectives | 5 | 0 | 0 | 1 | Hays asks if current physical robotics pose near-term existential risks, which Coogan counters by noting that safety proponents worry about future agentic superintelligence rather than narrow robotic automation. | |
| Bernie Sanders' Legislative Push and Industry Pushback | 6 | 0 | 0 | 0 | The hosts review Bernie Sanders' proposed legislation introducing 20-year prison terms and the 'corporate death penalty' for AI development. Coogan reviews industry responses from Fluidstack and other frontier builders. | |
| Cultural Memes, Corporate Satire, and Ecosystem Divides | 7 | 2 | 5 | 5 | Tyler Cosgrove engages directly with Coogan on Tyler Cowen's prediction market challenge. Cosgrove argues for falsifiability using the concrete example of nuclear bunkers, while Coogan pushes back using nuclear treaty history and deconstructs Paul Christiano's portfolio. | |
| Simulating Fruit Fly Brains and Moral Questions of AI | 5 | 3 | 1 | 1 | Cosgrove briefs the hosts on Google's full connectome mapping of fruit fly neurons. The conversation shifts into a light philosophical inquiry into the morality of simulating and tormenting virtual minds. | |
| Personal AI Workflows: Gaming in Balatro and Sim Racing Coaching | 4 | 0 | 0 | 0 | Coogan and Hays wrap up by sharing anecdotal use cases of computer use playing Balatro and ChatGPT coaching telemetry in iRacing at Laguna Seca. |
Statements from this episode (14)
Coogan: Congress is waking up to AI doomsday threat
“Congress is suddenly waking up to the AI doomsday threat.”
Coogan: Compute inspections in US data centers are very easy to enforce
“Very easy to enforce, at least in the United States.”
Coogan predicts chip controls and datacenter slowdowns are the likeliest AI brakes
“The big hammer is just chip controls and data center build out, slow down. That's the easiest thing, and I think that's, like, the biggest valve that they're going to be, that we're going to see twisted around to actually slow down capabilities.”
Hays: Strict AI regulation would incentivize millions to pursue secret breakthroughs
“And one thing that feels obvious around this if these policies were to be rolled out is that you would effectively create an incentive for millions of Individuals or groups globally to be in secret, like trying to find entirely new breakthroughs that are.”
Coogan: Ilya Sutskever's SSI recently received a large NVIDIA compute cluster
“SSI, Ilya Sutskiver's new, Neolab is recently got a big cluster from NVIDIA”
Coogan: AI clusters will be trackable from space via heat signatures
“Everyone is sort of with the consensus that it's going to be a big building with a lot of energy, big heat signature, definitely visible from space and pretty simple to track at least in the short term until people start building crazy underground facilities.”
Coogan: Debating exact p(doom) percentages and extinction mechanics is a sideshow
“The conversation definitely gets dragged down into PDOOM estimates and trying to narrow down exactly how a human extinction scenario plays out, and that can be that can, I feel like that's almost a sideshow. Because in a democratic society, in, amongst humanit…”
Coogan: Panic over AI slowdowns is unconvincing given existing capability overhang
“For most of these slowdown proposals, I personally have a hard time blackpilling about them, in the sense of, like, if the, if all of this gets implemented, how frustrated will I be? Like, the models are good. I would like better models. I want safe models, bu…”
Coogan: Current AI robotics poses no safety risk without model volition
“I don't think we're at a point where that poses a risk. It's the next model. It's the model with its own volition, basically.”
Coogan: Sanders' definition of superintelligence matches goals of 17 AI companies
“It sounds like the mission statement. It sounds like the explicit goal of like, 17 different companies right now.”
Coogan predicts pushback against AI regulation that favors top ten incumbents
“You're going to see a lot of pushback from people who are like the regulatory stuff is going to be like these 10 companies and I'm going to be number 11 and I'm basically getting the corporate death penalty then because I didn't make the cut to be one of the r…”
Coogan: AI doom concerns are largely contained to frontier labs
“The doom is very much contained to the frontier lab work. Everyone in the application layer who's applying the models, diffusing them, they're like, I just, I can't get this thing to work right. I gotta get forward deployed engineers to teach people how to use…”
Coogan: Lacking a bunker does not disprove belief in AI doom
“So if you think it's so totalizing, then you don't make the bunker. And so saying, hey, you don't have a bunker is not proof that the person doesn't believe what they're saying.”
Coogan: Paul Christiano's portfolio bets on AI success, not doom
“This doesn't seem like a doom based bet. This seems like this bet also pays out in just like the good ending and like AI. Is real and delivers value. So your AI bets perform well, the market performs well, and money slides from US debt to data centers and AI b…”