Nov 7, 2026 · 1h 47m
Politics collides with AI’s race for compute
The Great Collision Between Politics and AI | #309
The episode asks whether elections, regulation, energy limits, and public resistance can constrain AI before technology reshapes those institutions.
- 1Political resistance to data centers could raise AI’s costs, slow breakthroughs, and strengthen the largest incumbents.
- 2The hosts argue that recursive self-improvement may already be accelerating AI research, even as its measurement remains contested.
- 3Orbital and distributed infrastructure could eventually move critical compute beyond local political control, creating new governance problems.
Don't miss
The hosts interpret long-exposure images filled with satellite trails as an early visible stage of a Dyson swarm and debate the governance of orbital compute.
The brief
Anthropic’s possible IPO becomes an opening case study in AI’s capital needs: election uncertainty, valuation pressure, and the urgency of financing an expensive compute race.
The central political question is whether voters oppose AI itself or resent subsidizing data centers, power demand, and infrastructure that may mainly benefit large technology companies.
The proposed escape routes range from self-powered, underwater, and underground facilities to Canadian and orbital data centers, potentially shifting compute beyond municipal control.
The hosts argue that recursive self-improvement may already be underway as AI systems accelerate research, while benchmarks struggle to capture progress across tokens, watts, dollars, and economic output.
The discussion ends in space: satellite trails become a visible early sign of expanding orbital infrastructure, raising questions about congestion, ownership, and governance beyond Earth.
What was said on this episode
50 statements · 40 positive · 5 negative · 1 mixed · 4 neutral
Anthropic is unlikely to delay its IPO
“I really doubt they'll delay”
Listen at 0:22
Investment bankers are pressuring Anthropic to lower its IPO valuation
“investment bankers are trying to force down the IPO valuation”
Listen at 0:44
Meta became an exceptionally attractive buy after its IPO selloff
“it was the best buy in the history of the world if you just waited out a few weeks after the IPO”
Listen at 1:31
Meta’s value rose roughly 20–30 times from its IPO-era level
“it's up, what, 20, 30x, something like that”
Listen at 1:46
Regulatory and legislative headwinds facing Anthropic and OpenAI will eventually pass
“this too shall pass”
Listen at 3:02
Successfully managing a thousand concurrent AI agents could create dominance
“if you crack the code on that, I mean, you can be just dominant”
Listen at 4:43
Data centers will increasingly move into space
“the data centers are going to space”
Listen at 7:58
Terrestrial uncertainty will incentivize moving compute into space
“an incentive to start to push compute out to space”
Listen at 8:30
Frontier labs should solve the top 5,000 diseases in silico within two years
“solve the top 5,000 diseases, do it in the next one or two years, at least in silico”
Listen at 9:30
Frontier labs should use superintelligence to solve and market cures for 5,000 diseases
“Use superintelligence, solve the top 5,000 diseases, market the heck out of it to the general public”
Listen at 10:06
Municipalities should use their leverage in negotiating with data centers
“what they should be doing is recognizing that for the moment, they have leverage”
Listen at 18:04
Local governments should welcome data centers while demanding concessions
“don't push them away. Welcome them with open arms, but recognize that there could be a trade-off”
Listen at 18:28
Canada has substantially better per-capita power supply than other North American locations
“the power supply is overwhelmingly better per capita in Canada than it is anywhere else in North America”
Listen at 21:29
New Canadian data-center projects will likely come online around the next election
“a project that starts in Canada today will probably come online right around the time of the next election”
Listen at 21:48
Russia is moving some data centers underground because of the Ukraine war
“Russia is starting to put their data centers underground due to the war with Ukraine”
Listen at 23:41
Radical physics advances may substantially reduce data-center footprints within years
“we may achieve in the next few years radical advances in physics and applied physics”
Listen at 24:14
Markets priced in slower AI construction and an advantage for China
“the market just priced in. Yeah, this is going to slow things down and give China a leg up and a chance to bypass America”
Listen at 27:32
Space-based compute may represent only a few percent of total compute within two years
“best case, that's a couple percent of all compute in two years”
Listen at 27:50
Artificial superintelligence will arrive within two years
“we all know that ASI is within that two-year timeline”
Listen at 27:57
Compute constraints may make AI more expensive and reduce near-term breakthroughs
“the short-term effect might be more expensive AI for everybody and, you know, fewer breakthroughs”
Listen at 28:40
Political, financial, and social institutions are advancing slower than AI
“our political, our financial, our social institutions are sublinear”
Listen at 31:25
Software improvements may achieve superintelligence without new data centers
“through pure software improvements, get to where we're going, probably without any new data centers”
Listen at 34:13
AI products may substantially change the electorate within two years
“we're likely to hit. in the two-year timeframe, maybe even less”
Listen at 35:15
A free AI health app could materially improve health and build support for compute
“If you gave everybody a health app and let them track their health and give them material improvements with a free diagnosis”
Listen at 37:04
AI is already deeply engaged in recursive self-improvement
“We are so deep in recursive self-improvement at this point”
Listen at 40:32
AI has reached escape velocity toward superintelligence
“we're at superintelligence escape velocity at this point”
Listen at 41:38
Anthropic’s internal-development token use increased 235-fold in six months
“the use of tokens for internal development in Anthropic is up 235x in the last six months”
Listen at 42:12
Machines now conduct substantially all AI research rather than human researchers
“substantially all research at this point is actually being conducted by computer by machine researchers rather than by human researchers”
Listen at 44:51
“you can just RL or post-train your way from one in 10 to nine out of 10 ideas being good ideas”
Listen at 47:44
Most people will encounter AI directly for the first time soon
“their very first contact with AI should emerge very soon”
Listen at 50:58
A future optimal AI architecture will remain understandable to humans
“I don't think that it's going to be the case that when we get, assuming we get to a perfect architecture, that we'll look at it and we'll be scratching our heads”
Listen at 53:49
The optimal AI architecture will be beautiful and human-understandable
“I predict, it will be impossibly beautiful, so beautiful that even meat-body humans can understand how and why it works”
Listen at 54:14
Frontier AI models will shift to continuous release by early 2027
“I would expect that to happen early 2027 or so”
Listen at 1:00:16
Humanity may send relativistic von Neumann probes within a few years
“we're within a few years of having the capability of sending relativistic von Neumann probes in every direction”
Listen at 1:01:33
Human-centered podcasts and entertainment will persist despite AI assistance
“the podcast genre will go on and the human component of it will go on”
Listen at 1:02:59
Chinese open-weight models outperform American open-weight models in productivity per watt
“the Chinese open weight models are beating the American open weight models in terms of economic productivity per watt”
Listen at 1:06:17
Reversible computing could make Landauer’s limit economically irrelevant
“reversible computing could make the land hour limit effectively obsolete”
Listen at 1:07:28
Anthropic’s compute footprint may reach roughly four gigawatts by year-end
“Anthropic's total compute footprint by the end of this year at, call it four gigawatts-ish”
Listen at 1:14:27
Anthropic’s compute footprint may reach 16–18 gigawatts by 2027–2028
“potentially by 2027 in the 16 to 18 gigawatt range in 2027, 2028”
Listen at 1:14:40
Anthropic’s work will create immeasurable value for the world
“their endeavor will add so much value to the world, it's immeasurable”
Listen at 1:16:37
Entrepreneurial America needs a recognized right to access compute
“you have to have some kind of a right to compute”
Listen at 1:18:52
Restricted compute supply will increase compute prices
“one would expect the price of compute to go up”
Listen at 1:21:07
AI access could give businesses a 30–50% operating-margin advantage
“the neighboring bakery that does have access to AI will operate with at least a 30, 40, 50 percent operating margin advantage”
Listen at 1:23:13
Physical DNA and RNA synthesis should be regulated rather than AI thoughts
“I'd much rather that big brother be controlling the printing out of dna and rna sequences through tighter controls on a limited set of vendors than the safety cartel”
Listen at 1:29:45
Governments will likely regulate general-purpose robots similarly to biological synthesis
“I think we probably will”
Listen at 1:32:21
The deployment of orbital infrastructure marks the beginning of a Dyson swarm
“we're seeing the launch of the Dyson swarm”
Listen at 1:36:05
Earth’s skies already contain substantial orbital computing infrastructure
“Our skies are already filled with compute”
Listen at 1:37:13
Humanity will build a sun-synchronous orbital compute halo around Earth
“We'll build out our sun-synchronous orbit halo of compute around the Earth”
Listen at 1:38:20
Sun-synchronous orbit can accommodate substantial computing infrastructure
“We can put a lot of compute in sun-synchronous orbit”
Listen at 1:39:50
The long-term trend toward superintelligence will not reverse soon
“I don't foresee that secular trend reversing anytime soon”
Listen at 1:44:08
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
Some links are affiliate links — PodLume may earn a commission if you buy.
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.