Progress (and flying cities) via mining the discard pile

16.11, Friday 4 Aug 2023

Sometime around the 3300s, entire cities fly faster than light across the galaxy, working for hire for industrial planets. Populated by almost-immortals, powered and shielded by an antigravity device called a spindizzy the greatest of these is New York.

The spindizzy is more formally the Dillon-Wagoner gravitron polarity generator, built on the principles of the Blackett-Dirac equations: Every culture has its characteristic mathematic, in which its toriographers can see its inevitable social form. – from its first invention in 2018 (it was re-discovered later) the spindizzy enabled and ordained the takeoff and subsequent galactic society of itinerant cities.

The first book of James Blish’s Cities in Flight (They Shall Have Stars, 1956) covers the discovery of the spindizzy.

The West has entered a stagnant, paranoid state. Science has stalled. The scientific method has become obsolete: the more subtle the facts to be discovered become … the more expensive and time-consuming it is to investigate them – which leaves the unimaginative government as the only potential funder.

Instead? Dig through the slush pile, the crackpotssports, freaks, near-misses.

You need marginal contributors, scientists of good reputation generally whose obsessions don’t strike fire with the other members of their profession. Like the Crehore atom, or old Ehrenhaft’s theory of magnetic currents, or the Milne cosmology

So they dig for two years…

checking patents that had been granted but not sequestered, published scientific papers containing suggestions other scientists had decided not to explore, articles in the lay press about incipient miracles which hadn’t come off, science-fiction stories by practicing scientists, anything and everything that might lead somewhere.

And eventually find the Blackett equation, which had been suggested way back in 1948 but was not testable – at the time.

Then they test it, and it works, and they invent the spindizzy, and then they fly to Jupiter in 3 days, and a thousand years later Earth is a garden planet with its old factories and cities alike now aloft, flitting between the stars.

I love stories of the exhilaration and work of scientific discovery. I could read them all day.

Anyway so X-nee-Twitter is alive with talk of LK-99, the Ambient Superconductor of the Summer (New York Times), which purports to be a miracle material, discovered in South Korea, unveiled to the world with much surprise and fuss, and - if real - is a capital-B Big capital-D Deal in that room temperature superconductivity would enable tiny electronics, maglev trains, and all the rest.

(Honestly we could just build regular trains without superconductor magnet levitation and that would be great too, but that’s another story.)

Of course it’s extremely likely that the LK-99 claims are not true. The papers on arXiv are missing the actual evidence of superconductivity, the labs trying to replicate the claimed characteristics of the material have not yet been successful, and the news is somewhat suspiciously timed given that (I hear) the material itself was discovered in 2020 and its disclosure now coincides with the company behind it trying to raise money.


I have a fantasy of a glitch in the universe. And so I am happy to bathe contented in the possibility and enjoy these few days before the fantasy comes crashing down, just as it did with desktop cold fusion, and just as it did with Italian FTL neutrinos.

Which is why I went back and re-read the beginning of Cities in Flight.

There’s something about Blish’s idea of winnowing the chaff of the scientific discard pile.

Like: maybe new ideas are ten a penny, and the scientific progress pipeline is replication-bottlenecked.

It reminds me of writing. There’s drafting and editing and the two are distinct. Perhaps, in science, we don’t need new ideas (drafting) right now. But editing is constrained.

For two reasons. Being (a) there aren’t enough people who can follow up on ideas, and (b) ideas are put forward and its not at the time possible to replicate or test them, or they don’t fit into the conceptual framework… yet.

I’m thinking of VR?

The problem with virtual reality, which has been tried a number of times over the decades, isn’t that it’s a fundamentally bad idea that makes 20% of people VR sick. It is eye-wateringly astounding even if I do have to chew ginger gummies to keep the nausea down when I’m wearing my headset.

No the hurdles have been miniaturised technology (screens and gyros) and go-to-market strategy, Apple’s take on the latter being social impedance matching (external screens so you can still make “eye” contact) and positioning the first read of the Vision Pro product as basically a large, high DPI external monitor that also happens to be portable, and also also happens to enable multiplayer VR (but that’s not the initial sell), i.e. a bargain.

These can be overcome with cash and will, sure, but as challenges they’re not in the same taxon that impeded AI, for example, which required a true and fundamental breakthrough.

Maybe it is worth going back to other ideas from the 1980s and seeing what else didn’t work then, being just too wild, but would today, given four decades of technology, industrial base, and consumer readiness?

Like, what’s the go-to-market strategy for a Drexler molecular assembler?

This is just for example but what I’m saying is, could we have another run at 3D printers? Think really, really hard about the consumer go-to-market. Maybe we missed something last time round; maybe the context has changed over the years.

I would automate this abandoned invention threshing process.

  • First I would train an AI on industrial processes, giving it access to Alibaba, some manufactured goods simulation software, and arXiv for novel material processes.
  • Then I would turn it loose on the Technovelgy database of inventions in science fiction. For example, looking at the inventions by Bruce Sterling, how about an air-conditioned coat, or a self-actualising stone soup Bambakias Hotel?
  • Finally, if there’s anything technologically feasible, ask the AI to come up with a scatter of positioning statements, auto-purchase ads on Instagram to match, and see what gets clickthrough.

Then scale what works.

RELATED: I also have an essay on the automated discovery of new areas of thought, which starts with the concept of horsehistory.


What ideas did I once upon a time flip the bit on as unachievable but actually I should now revisit? That’s the real lesson here.

For example: back in 2021 I had an idea for a galactic compass, that is, an iPhone app with a floating arrow that always points directly to black hole at the middle of the Milky Way.

I couldn’t build it then. But recently I realised that ChatGPT could walk me through writing an iOS app. So now it’s an actual thing! On my homescreen! It works! (Mostly.)

(Hey lmk if you’d like to get on the TestFlight as a beta user, I just need your Apple ID. Also let me know if you understand a bit about SceneKit and combining rotations with quaternions, because I’m in a muddle with the math and I’m getting some drift on the azimuth. The astro equations are fine, but I’ve got a 10 degree error in combining it with the device rotation. I need help fixing that bug because GPT, not being embodied, is even worse than I am at thinking in 3D.)

But not just tech, right? All kinds of personal ideas.

What do I deep-down believe is out of reach for me that I could now achieve if I just tried? One’s mind bends around assumed-impossibilities; I can’t imagine them to enumerate them. I wonder what the threshing algorithm would be like in order to dig them up.

If you enjoyed this post, please consider sharing it by email or on social media. Here’s the link. Thanks, —Matt.