Jan 31, 2024 · 39m · american-optimist
Ep 76: Accelerate or Die! Guillaume Verdon and the e/acc Movement · Joe Lonsdale
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 American Optimist, host Joe Lonsdale interviews theoretical physicist and effective accelerationism (e/acc) founder Guillaume Verdon to explore how thermodynamic physics, techno-optimism, and permissionless innovation can drive civilizational progress. Together, they discuss overcoming institutional resistance, unlocking nuclear energy abundance, and leveraging open-source AI and thermodynamic computing to secure a prosperous future.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Joe holds 25.1% of the talking time here. How this is scored →
Joe as informed peer 4.2 Guest teaching 4.8 Guest disagreement 1.3 Joe pushing back 1.2
speaking balance: gold is Joe, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 16:13 Rejecting Authoritarian and Viral Stability
Verdon counters the premise that pure optimization favors authoritarian empires or parasitic viruses, arguing that oppressive systems inherently self-destruct and get outcompeted.
Hardest push from Joe ▶ 17:31 Devil's Advocate on Morality and Human DignityLonsdale directly challenges the moral adequacy of pure competence-driven growth by raising Judeo-Christian obligations to protect the vulnerable bottom decile.
Biggest teaching moment ▶ 4:06 Physics of Holographic Information CompressionVerdon educates the host on how black hole surface area entropy in theoretical physics directly inspired new algorithms for quantum deep learning.
Joe holds their own ▶ 37:59 Lonsdale Details Institutional Capital in D.C.Lonsdale demonstrates his deep political and institutional footprint by outlining multi-state operations and venture coalitions backing open-source advocacy in Washington.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Joe as informed peer | Guest teaching | Guest disagreement | Joe pushing back | Why |
|---|---|---|---|---|---|---|
| Monologue: The Cultural Importance of Technological Acceleration | 3 | 6 | 0 | 0 | Joe introduces the guest and provides context before prompting Guillaume on his academic journey. Guillaume breaks down advanced concepts linking theoretical physics, Claude Shannon's information theory, black hole entropy, and quantum deep learning. | |
| Aligning with the West and Cultural Dynamics at Google | 4 | 5 | 1 | 1 | Joe asks about geopolitical alignment and practical quantum technologies. Guillaume explains quantum sensing and navigation applications that solve GPS vulnerability in conflict scenarios. | |
| Institutional Resistance and Cultural Engineering Roots | 3 | 4 | 2 | 1 | Joe inquires about Google's internal reluctance to collaborate with the Department of Defense. Guillaume details the cultural dynamics inside big tech that motivated him to create e/acc and engage in cultural engineering. | |
| Thermodynamic Foundations of Effective Accelerationism | 5 | 6 | 2 | 4 | Guillaume outlines the thermodynamic and evolutionary roots of e/acc. Joe plays devil's advocate, challenging whether unconstrained system growth resembles parasitic or authoritarian expansion. | |
| Human Dignity, Moral Frameworks, and Cultural Diversity | 6 | 7 | 2 | 5 | Joe presses Guillaume on human dignity and Judeo-Christian morality versus ruthless Nietzschean optimization. Guillaume frames religious and moral codes as evolutionary heuristics that maintain societal variance and long-term robustness. | |
| Nuclear Energy Abundance and Human Population Scale | 4 | 4 | 1 | 1 | Both agree strongly on expanding nuclear energy and removing regulatory barriers to support massive population scale and technological abundance. | |
| Hyperstition: How Culture Shapes Future Reality | 4 | 3 | 1 | 0 | Guillaume explains 'hyperstition'—the idea that narratives and cultural expectations actively manifest reality. Joe validates this from his entrepreneurial career building multi-billion dollar firms. | |
| Overcoming Pessimism and Incumbent Monopolization | 4 | 4 | 2 | 0 | Joe and Guillaume discuss turning cultural pessimism around and preventing corporate incumbents from using fear-mongering to entrench regulatory monopolies over AI. | |
| Historical Resistance to Technological Innovation | 4 | 6 | 1 | 0 | Guillaume introduces his new venture Extropic and explains thermodynamic computing, which leverages physical thermal fluctuations to accelerate generative AI computing efficiency. | |
| Scaling e/acc and Protecting Open Source AI in Washington | 5 | 3 | 1 | 0 | The conversation concludes on policy strategy in Washington D.C. to defend open-source AI. Joe highlights his venture and advocacy networks actively lobbying for these principles. |
Statements from this episode (23)
Verdon: He pioneered the field of quantum deep learning
“I was basically a pioneer of a field called quantum deep learning. So many people are familiar with deep learning neural networks, but extending those types of notions to quantum mechanical computers and really defining what does it even mean to have a sort of…”
Google acquired Verdon's quantum TensorFlow prototype and hired his team
“During that first startup I ended up Being approached by Google to prototype, really, what would a TensorFlow for quantum computing look like? So TensorFlow at the time was the core software framework for Google to do deep learning, and we wanted to extend tha…”
Verdon: Quantum Sensors and Communication Are Much Shorter-Term Than Quantum Computing
“I would say quantum sensors and quantum communication are much shorter term technology than quantum computing, and that's why I moved towards those fields after subfields after a couple years in quantum computing.”
Verdon: Quantum Sensor Prototypes Outperform Classical Sensors on Positioning and Timing
“I do think there are very solid prototypes that can resolve regions of parameter space in terms of acceleration positioning and timing. That are not possible with classical sensors. And now it's a question of miniaturizing and robustifying those systems.”
Verdon: Big Tech ideological subversion stops government collaboration, helping adversaries
“What I saw with my own eyes was that a sort of cultural subversion within big tech more broadly was stopping a big tech from working with the government, right? Like basically ideologies that sort of were very pervasive within the middle layers of the organiza…”
Verdon: Deploying advanced technology requires active cultural engineering
“Even if you can work on technologies all day, but those that get deployed and funded you know, if you don't have a sort of ideological or cultural support for what you're working on you're not gonna get to build and deploy certain technologies. And that kind o…”
Verdon: Started e/acc via anonymous Twitter alt at Google X
“I think I think I started it roughly when I was towards the tail end of my stint at Google X there, obviously the baseline is secrecy. I couldn't talk much and you know, about what I do and can't, couldn't say anything too controversial. And so, you know, I st…”
Verdon: Societies must maintain startup-like adaptability to maximize growth fitness
“And so one of the sort of core tenets is to maintain variance, maintain adaptability, because, you know, it's just like an early stage startup versus a late stage company. An early stage startup, things aren't fully pinned down. You're still flexible. You're s…”
Verdon: Physics predicts civilization optimizes to consume more energy
“At least from the equations of physics, if you consider the whole system of civilization, the equations predict that the system will optimize towards its growth, towards consuming more free energy.”
Verdon: Over-regulation, not physics, is holding back nuclear energy
“Definitely seeking forms of energy that are more potent, such as nuclear energy which right now it's not necessarily the physics that's a problems over-regulation.”
Lonsdale: Earth could comfortably support 100 billion people
“I feel like with nuclear energy and things like tunnels around our major cities and stuff to get rid of traffic entirely, cause you can have like arbitrarily, you know, large amounts of going in and out. It seems like there's no reason we couldn't have a hundr…”
Verdon: Market incentives and liability will ensure AI alignment
“People wouldn't pay for an AI that's unaligned, doesn't do what you say, and is nefarious, right? They would pay for an AI that does, you know, is a good worker, let's say is aligned and is reliable, right? And if the, you know, the person or the company that …”
Verdon: Multi-year regulatory cycles are too slow for AI progress
“AI is moving on a timescale of month to month, week to week, and, you know, regulations are on a few-year timescale at best put together, and that's just too slow of a feedback loop, and it's too heavy-handed, and it's probably not the way forward.”
Lonsdale: Regulations make US healthcare three times as expensive
“It's the same thing in healthcare in America, all the incumbents, right? All the regulations. And so you can't adapt and the healthcare is like three times as expensive or something as it would be.”
Verdon: Centralized bioweapon defense research accidentally created a damaging virus
“There was centralized research on Bio, you know, bioweapon defense. Right. And we were worried about a potential virus that could do a lot of damage. And then it was actually accidentally created through that research.”
Verdon: Preemptive over-regulation does more harm than technological risks
“We can preemptively do much more damage to ourselves with, let's say, over-regulation than You know, the damage that would have happened by just letting things happen.”
Verdon: Incumbents use fearmongering to consolidate power and entrench dominance
“Fear mongering becomes instrumental towards certain people and incumbents that are trying to gather power and crystallize their position before, you know people realize that actually the future wasn't going to be a gloom and act actually what is going to be gr…”
Verdon: Extropic is developing self-adaptive hardware with positive global impact
“We're working on very sophisticated, you know, self-adaptive hardware that could scare some people, but, you know, we think it's going to really impact the world positively.”
Verdon: Extropic thermodynamic compute will reach production in a couple years
“I'm, I haven't announced any prototypes yet, but you know, we have our first test chips, and we're iterating, but I think a couple years from now, we'll be able to have this in, in production.”
Lonsdale: Has teams in 18 states and hired D.C. advocates
“I have teams in 18 states, and I just hired some people who are great in D.C. Just for purposes like these.”
Lonsdale: Marc Andreessen is spending $75M on D.C. tech policy lobbying
“I know Marc Andreessen and his team spending seventy-five million dollars for the first time hiring people in D.C. To fight for these values”
Verdon: Robotics Will Boom Within a Few Years to Solve Blue-Collar Intelligence
“I think robotics is gonna see a boom in, in the coming years as we, you know, cross that threshold of having sort of you know, match, being able to have synthetic white collar intelligence. I think there's a few year gap between we'll be, Until we can solve mo…”
Verdon: Operational AI Leverage Will Cause Deflation in Real-World Construction Costs
“You have intellectual leverage now with Chan GPT, but eventually you'll have operational leverage, and that's where we're gonna see serious progress in the world of atoms. We're gonna see deflation and the cost of building things in the real world”