Between 1991 and roughly 2006, a few hundred people argued with each other over email.
The list was called Extropians. It ran out of the Extropy Institute, a small California nonprofit founded by Max More, then a philosophy PhD student at USC, and Tom Morrow. The name came from a coinage: entropy is the tendency of a system toward disorder, so extropy was the informal opposite, a measure of a system's intelligence, order, vitality and capacity to grow. The whole project was the belief that you could increase it on purpose.
The subscriber count peaked in the hundreds. Traffic ran to fifty or a hundred messages a day. By any reasonable metric it was a niche hobbyist forum for people interested in cryonics and space colonies.
Here is what came out of it.
Wei Dai published b-money to the list in 1998: a currency with no issuer, where anonymous nodes maintain a shared ledger and computational work creates the coins. Nick Szabo was working on bit gold at the same time, chaining unforgeable proof-of-work strings so scarcity could be proven without a mint. Neither was ever built. Both are cited in Satoshi's whitepaper. Hal Finney, another regular, built Reusable Proof of Work in 2004, then ran the second Bitcoin node and received the first transaction in 2009.
Szabo also coined smart contracts on that list in 1994, which is the direct ancestor of Ethereum and everything downstream of it.
Robin Hanson worked out prediction markets in those threads, the idea that if you let people bet on factual claims the price becomes the best public estimate of truth. That is now Polymarket, Kalshi, Metaculus, and a standard tool in forecasting.
Eliezer Yudkowsky joined as a teenager and spent years arguing about what happens when a machine can improve itself. That argument became the Singularity Institute in 2000, then MIRI, then LessWrong, and from there the alignment teams at every major AI lab and the UK and US AI Safety Institutes.
Nick Bostrom was there. He went on to found the Future of Humanity Institute at Oxford and give existential risk its modern shape.
One mailing list. Three or four hundred people. Bitcoin's intellectual lineage, prediction markets, AI safety as a field, and the longtermist wing of effective altruism. I cannot think of a better return on a mailing list in history.
What made it work
I have spent a while trying to figure out why that room produced so much, because I would like to reproduce it.
A few things stand out.
It was small enough that reputation was real. Everyone read everything. If you made a sloppy argument, four people who had thought about it longer than you took it apart in public, and you had to answer them. There was no way to be vaguely impressive.
Nobody had permission. There was no grant to lose, no department to embarrass, no journal deciding what counted. Drexler was writing about molecular assemblers, Hanson about betting markets for scientific claims, a teenager about recursive self-improvement. In 1995 all of those were career suicide in any institution. On the list they were Tuesday.
The ideas were allowed to be linked. That is the part people miss. The list did not have a cryonics thread and a separate crypto thread. It had one long argument in which life extension, digital cash, nanotechnology, AI, and private law were all facets of the same question: what happens when you take the constraints on human life to be engineering problems rather than facts.
And it was written. Everything was text, archived, quotable, attackable. b-money exists because Wei Dai wrote it down where Szabo and Finney could read it.
The blind spot
Now the uncomfortable part, and the reason I am writing this.
Look again at what actually shipped from that list.
Bitcoin. Smart contracts. Prediction markets. AI safety. Existential risk philosophy.
All of it is information. Every single one of the Extropian ideas that became real is made of bits. Not one of them required moving an atom.
And the one big physical bet the list made, the one thing that would have changed the material world rather than the informational one, is the one that failed.
Molecular nanotechnology was central to Extropian thinking. Eric Drexler's Engines of Creation was near-canonical on that list. The promise was machines that position individual atoms, enabling near-free manufacture of almost anything. In 2001 Richard Smalley, who had a Nobel for discovering fullerenes, publicly attacked the concept in Scientific American. The exchange ran through 2003 in Chemical & Engineering News. Smalley's "fat fingers and sticky fingers" objection, roughly that a molecular assembler cannot both grip and release atoms with the precision required, was widely read as a knockout. Funding moved. The field renamed itself, folded into materials science, and the original vision went quiet for twenty years.
I think that is the single most consequential thing that happened to the Extropians, and almost nobody talks about it that way. The list's descendants inherited a bias toward the layer where they had won. Crypto and AI are cheap to iterate. You can test a protocol on a laptop. Matter is expensive, slow, and regulated, and it punishes you for being wrong.
So a generation of extremely capable people who believed civilization was an engineering problem spent thirty years engineering the parts of it made of information, and left the substrate alone.
Meanwhile the substrate is what is actually failing.
What the new list is about
Everything I work on now sits under something I call Heliogenesis, from helios and genesis, a beginning organised around the sun. The core number is simple: the sun delivers roughly 173,000 terawatts to Earth's surface. Human civilization currently runs on about 18. Everything we fight over is inside a rounding error of what arrives for free every day.
The gap is not an energy problem. It is a conversion and materials problem, and it lives entirely in the layer the Extropians skipped.
Three things have changed since 2003 that make me think the physical layer is now where the leverage is.
Molecular manufacturing came back, quietly. My friend Jeremy Barton runs the NanoDynamics Institute and has spent fifteen years on atomically precise fabrication. Scanning probe fabrication, molecular machine control, multi-axis electronic control systems for molecular machines. The near-term goal is unglamorous and enormous: basic molecular motors, then the first functional nanomachines, then machines that build machines at that scale. Desktop fabrication of advanced materials from minimal energy and feedstock. Smalley won the argument in 2003 against a version of the idea that was largely a thought experiment. Barton is running experiments. That is a different situation, and Phase 1 of his roadmap costs about the same as a mid-size seed round.
Minerals turned out to be the binding constraint. Simon Michaux at the Geological Survey of Finland has done the arithmetic on what a full transition to renewables plus electric transport would require in copper, lithium, nickel, cobalt and rare earths. The numbers do not close against known reserves. He calls it minerals blindness. Almost every green transition plan assumes materials are available and priced, which is the same assumption the Extropians made about the whole physical world.
Biology became a manufacturing method rather than a subject. Anne Meyer at Rochester is 3D printing living materials and producing biomaterials bacterially. Plantd is making carbon-negative structural materials out of giant reed. Zsofia Kollar at Human Material Loop turns waste human hair into keratin fibre for textiles and insulation, using two processing steps and zero land. None of this existed as a practical option when the Extropians were arguing.
Those four things belong in one conversation and are currently in four. That is exactly the condition the Extropian list solved, and the reason I think a list is the right instrument rather than a conference or a fund.
The other thing that has to change
There is a second lesson from the Extropians, and it is about ownership.
b-money and bit gold were published, not patented. That is the only reason Bitcoin exists. If Dai and Szabo had filed, the most important monetary innovation in a century would have died in a portfolio somewhere.
That happened by accident. It happened because those two were writing to a mailing list rather than to a VC, and the norm of the room was to publish. Nobody designed it.
The Extropians were also relentlessly individualist. Self-direction and self-transformation are two of the seven Extropian Principles, and the political frame was anarcho-capitalist by default. It produced beautiful protocols and no answer at all to the question of how a civilization coordinates.
If a room like this comes together again, the openness has to be deliberate and the unit of concern has to be larger than the individual. In the Oslo Project work we have been writing this into partnership agreements: research funded unconditionally so that the pressure to patent disappears, breakthroughs held in a commons, defensive patents only, deployment priority to Global South and Indigenous-led regions rather than to whoever can pay. Whether or not that specific structure survives contact with reality, the principle is the point. The Extropians got the commons by luck. We should not rely on luck twice.
So: a letter
I want to start a small list.
Not a newsletter in the growth sense. I already write publicly and that is a different thing, aimed outward, measured in subscribers. This would be aimed inward: forty or fifty people who are actually building at the physical layer, writing to each other in long form, in the open, with the expectation that ideas get taken apart.
The rough shape I have in mind:
Small and by invitation. The Extropian list worked at a few hundred and stopped working when it got bigger and moved to blogs. I would rather start at forty and be strict.
Written, archived, citable. A letter, not a chat. Chat produces vibes. Letters produce b-money.
One subject, broadly construed. The material basis of a civilization that can run on solar flux without eating its substrate. Energy conversion, molecular manufacturing, materials, biological fabrication, and the governance and capital structures that decide who gets any of it. Not AI, not crypto, not because they do not matter but because they already have their rooms.
Open by default. Anything of consequence gets published where anyone can build on it.
A working obligation. Not a reading group. Everyone on it should be building, funding, or governing something real, and should be willing to write about what is not working.
I do not have a name for it yet. I do not know whether the right cadence is weekly or monthly, or whether it should be a list, a private Substack, or something self-hosted. I have not decided whether it stays forty people or whether that is just cowardice dressed as curation.
What I do know is that the last time a room like this existed, three or four hundred people rewrote money, forecasting, and the entire discourse around machine intelligence, from a mailing list, with no money and no institutional cover, and it took about fifteen years to show up in the world.
The one thing they did not touch is the thing that is now breaking.
If you are working on any of this, write back and tell me what you are building and what you think the list should be for. I will send the first letter to whoever answers.