Can We Uninvent a Technology?

Making things secret again, even mysterious

(roughly ~7,400 words, estimated reading time 35-40 minutes @ 200 wpm)

Supplemental reading:

• Jorge Luis Borges, “The Library of Babel”

• Nick Bostrom, “Existential Risk Prevention as Global Priority”

• Frederick P. Brooks, The Mythical Man-Month

• Richard P. Feynman, Feynman Lectures on Physics

• Leonard E. Read, “I, Pencil”

• Thomas Thwaites, The Toaster Project or, a Heroic Attempt to Build a Simple Electric Appliance From Scratch

• Peter Thiel, From Zero to One: Notes on Startups or, How to Invent the Future

• Jack Vance, The Demon Princes (series)

---

“Break the Tools” generated by wombo.art

Imagine you have some serious reservations or just a persistent grudge against a particular technology: maybe thermonuclear weapons, maybe CRISPR-Cas9 gene editing, maybe the Bluetooth headset. Let’s further imagine your grudge extends beyond sanity into outright supervillainy: not only do you want people to stop using the technology, you want them to stop being able to recreate the technology. What you want is not just the loss of the technology (all the Bluetooth headsets thrown into the sea, for example) but for the technology to be unrecoverable (humanity forgets how to make Bluetooth devices, or that they even existed in the first place).

Can you uninvent a technology? Let’s find out!

To Uninvent

Specifically, to uninvent X, where X is a technology is to make X no longer accessible to contemporary humanity, where ‘accessible’ is a combination of:

(1) - the technology is known to be possible by at least some people (e.g. nuclear fission for weaponry in the mid-1930s, which was theoretically possible. What followed after some initial feasibility experiments - Fermi’s U of Chicago chain reactor using graphite and natural uranium, in particular - was engineering, not basic science).

(2) - following from (1), the construction process / diagrams / blueprints are known to at least some people (e.g. nuclear weaponry designs at Los Alamos before August 1945)

(3) - following from (1) and (2), can be manufactured with present techniques and equipment (even if the specific tools are not yet set up to do so (e.g. separating U-235 from U-238 at industrial scales to fuel a nuclear bomb; the techniques were known to be possible by 1943 though the extraction facilities were not yet set up).

To uninvent a technology would be to make (1) through (3) all false.

Apocalypse Maybe Not…

A restriction on this thought experiment: you do not have access to a time machine. You cannot go back and push the inventor in front of a bus before he has his “eureka!” moment. No mind control either: hypnotizing the world through television or streaming video services to forget about a technology while your agents remove all traces of it. In this thought experiment we can apply force, political persuasion, and money only.

A further restriction: no Gordian Knot cutting by collapsing technological civilization.

It’s a commonplace that our technological world is fragile, in the Nassim Taleb sense that it suffers from disorder - add enough disorder to the world system, and it will break down, like a teacup being shaken violently in a metal box. After a Global Catastrophic Event, such as a near miss by a rogue planet, a superplague that kills 90% of humanity, or a not-limited nuclear war between two superpowers, the survivors would be left with only the remnants of technology, with no ability to repair the few artifacts that still worked or to build new ones.¹

Destroying humanity or at least wrecking our technological civilization would suppress and definitely uninvent just about any technology that irks you (except possibly fire), but presumably you want to be rid of nuclear weapons/genetic engineering/Bluetooth while still being able to enjoy antibiotics, automobiles, and air conditioning. What to do then?

A Case Study

First, a real life example of how a technology was nearly lost.

The National Nuclear Security Administration (NNSA) is in charge of maintaining the United States’ Enduring Stockpile, the nuclear weapons retained from the end of the Cold War as a hedge against other nuclear weapons states. During the maintenance cycle for one of those weapons, the W76 warhead² used in Trident submarine-launched ballistic missiles, the engineers opening up the bombs discovered that they needed to replace a component, code named FOGBANK. The problem was, they were not sure how to make it. Despite the fact that the Y-12 National Security site in Tennessee had a building dedicated to making the material, knowledge of how to do so was largely in the heads of the chemists and engineers who worked there. All of them had retired, and they did not leave sufficient notes to restart the process.

Roughly $70 million and a lot of “please, please, please come back and tell us how you did this” contracts later, the NNSA was able to make FOGBANK again, the warheads were refurbished, and the second leg of the nuclear triad returned to sea. So much for genie out of the bottle myths - if they had waited a few more years, an entire class of nuclear weapons might have gone out of service prematurely.

Here we have a technology that was nearly lost because it was secret, difficult to make, and consequently known only to a very few people. Simple neglect led to a decay of institutional knowledge. Could we do the same thing for a technology that is not secret, widespread, and has multiple manufacturers?

The Act of Uninventing

Many, perhaps almost all, of the technologies we use each day, even the most commonplace ones like a toaster or a pencil, are dependent upon enormous amounts of distributed knowledge and other technologies. Leonard E. Read, in his classic essay ”I, Pencil,” points out that there is no person on Earth who knows every detail of how to make a pencil from raw materials to finished product. To start with, consider all the know-how that just goes into harvesting the tree out of which the pencil is made (in the essay, the pencil speaks in the first person):

My family tree begins with what in fact is a tree, a cedar of straight grain that grows in Northern California and Oregon. Now contemplate all the saws and trucks and rope and the countless other gear used in harvesting and carting the cedar logs to the railroad siding. Think of all the persons and the numberless skills that went into their fabrication: the mining of ore, the making of steel and its refinement into saws, axes, motors; the growing of hemp and bringing it through all the stages to heavy and strong rope; the logging camps with their beds and mess halls, the cookery and the raising of all the foods. Why, untold thousands of persons had a hand in every cup of coffee the loggers drink!

Leonard E. Reed, “I, Pencil”

Thomas Thwaites, a British art student, demonstrated the truth of Read's essay through his attempt to create a working toaster from raw materials, including gathering (and smelting) his own iron ore and mica. Spoiler: he did not succeed, but the journey, documented in his book The Toaster Project, is a wonderful read.³

To the would be uninventor, this should be encouraging. If you wanted to get rid of nuclear weapons or Bluetooth headsets, you only need to find a few critical nodes in the linkage of knowledge and industry, and you could shut those down. Maybe you could acquire the patents, or buy up the natural supplies, or just retire with nice pensions all the people with the know how before they could pass it down. In a generation or two, the technology would be lost.

What you would hope is that the ‘tech tree’ of a pencil looks like this:

Unfortunately, this won’t work even in the case of a pencil, because while specialized the knowledge for each stage and product is highly distributed and there is economic demand for it to be sustained. Designs for pencils, for the equipment to make the components, the chemical formulae to refine the raw materials, all of these are present in many libraries and archives. Hundreds of companies are present in each node of the tech tree. Biological trees are used for more than just pencils, and the metal in the saws that cut them have other applications beyond making saws. Even coffee has many other markets than logging camps.

The economic demand means that there are jobs for those who would acquire the skills and knowledge - chopping wood, metallurgy, coffee growing - necessary to make them, and so there is a ready supply of people to do the work. There also are no monopoly suppliers.

Even worse, the nodes are not clearly separable. One knowledge or industry node contributes to not just a node above itself, but to others out of its direct line to the pencil - metallurgy gets you not just the saw for the axe, but the steel frame for the logging truck, the shovel head with which the graphite is dug, and on and on. The tech tree of the pencil looks more like this.

You would hope this would be easier with more complex technologies, but the same factors are at play. Highly complex technology tends to be highly-profitable, and that drives even greater demand for specialization in its service. Even technologies with limited applications, like nuclear weapon components, have a lot of inertia behind their production. Whatever FOGBANK is (likely an aerogel, according to arms control experts)⁴, it was eventually remade, because a government with enormous resources wanted it done. Even the total loss of the knowledge of how to make it would have been recoverable, because they had ready to hand examples in the bombs themselves to reverse engineer. Or they could have worked out its function within the weapon, and created a substitute material

A Fictional Example

Jack Vance’s classic series of science fiction novels, The Demon Princes, takes place in the Oikumene, a collection of allied human worlds around the galactic centre. Despite being set several hundred years in the future, the only major science fiction technology on display is faster than light travel. Characters listen to music on tape decks, take and develop pictures on photosensitive film stock, place phone calls on landlines, and read printed newspapers, magazines, and books.

This all seems delightfully quaint, another example of how science fiction writers fail to predict the future, until you realize that it is deliberate. In the background, mentioned intermittently throughout the first three books, is a large, powerful organization called simply The Institute, to which many prominent citizens of the Oikumene belong. The Institutes purpose is not clear until the latter half of the series, when it is revealed that, by buying up patents, the Institute has suppressed technology it deems dangerous to social stability.

For instance, the citizens of the Oikumene used to have networked personal computers, but the Institute felt this was too dangerous so they bought up the patents and suppressed the technology. In the Oikumene, many technologies are known to be possible conceptually or from history, but the grip of the Institute - and the Oikumene’s common legal jurisdiction and infinite patent duration - make them essentially lost to humanity.

In our contemporary world, the best you can do is to hoard patents for their lifetime - around 20 years in most countries from the date of filing - and sit on the technology without licensing it, suing anyone who builds anything similar. Even this is most effective within single countries, or in jurisdictions like the European Union with a common legal framework. Even in the Oikumene, forbidden technologies are present in the periphery, the worlds beyond the settled ones, where there is no law beyond might makes right.

Searching in the Library of Babel

It is a profound and necessary truth that the deep things in science are not found because they are useful; they are found because it was possible to find them.

-Robert Oppenheimer

Technologies, like scientific knowledge, are hard to suppress. When people are motivated to search in the same space, with some idea what they are looking for, then, if it can be found, they will find it.

I like to use Borges’ marvellous story “The Library of Babel" to explain this. The story concerns an extensive, practically infinite library. The library is made up of interlocking hexagons, with shelves on five of the walls and passages between the hexagons accessible via the open sixth wall. No one has been able to count how many levels of hexagons there are.

Five shelves correspond to each one of the walls of each hexagon; each shelf contains thirty-two books of a uniform format; each book is made up of four hundred and ten pages; each page, of forty lines; each line, of some eighty black letters.

-Borges, “The Library of Babel”

Later in the story, it is specified that the books are written using only twenty-two characters, plus the space, period, and comma (n = 25). Further to that, a proof exists that there are no two identical books.⁵ The books are the complete permutations of the set of characters, spaces, periods, and commas. I leave the permutation analysis of the total number of books in the library as an exercise for the reader.⁶

The Library of Babel serves as a good model for the space of possibility. If an idea can be expressed in words or numbers, it is in the Library.⁷ Though welcomed as a gift to humanity upon its discovery, by the time the unnamed narrator is writing his manuscript the Library’s vastness has driven many of its librarians insane:

It is now four centuries since men have been wearying the hexagons…

There are official searchers, inquisitors. I have observed them carrying out their functions: they are always exhausted … From time to time they will pick up the nearest book and leaf through its pages, in search of infamous words. Obviously, no one expects to discover anything.

The uncommon hope was followed, naturally enough, by deep depression. The certainty that some shelf in some hexagon contained precious books and that these books were inaccessible seemed intolerable. A blasphemous sect suggested that all searches be given up and that men everywhere shuffle letters and symbols until they succeeded in composing, by means of improbable strokes of luck, the canonical books.

What’s maddening is that all possible books are in there, which means all the greatest works it is possible to write are there too. Though unspoken in the story, the reader infers the deeper truth behind Borges’ imaginative construction: even with the ability to generate the Library of Babel (a computationally impractical task)⁸, in order to find the book you want, you have to create it.

Creative minds are the search function to find the books.

What minds can access which books and bring them into our world (and here I bring this long metaphor back around to my point) is based on their position in spacetime. Homer could not have written Shakespeare’s plays, no more than Shakespeare could have written Homer. Artistic works, likes books, plays, and musical compositions, are singular: there exists only one, or at most a few variants that rise to the same aesthetic level.⁹

Truths about the natural world, or about technologies, are far more redundantly ‘encoded’ in the library. There’s only one true King Lear, and an enormous number of corrupted copies,¹⁰ the way there is only one true Well Tempered Clavier Book I, but there are more ways to state Newton’s Laws of Motion than just the way Newton put them in the Principia. Scientific and technological truths, which can be truth-preservingly expressed in multiple ways, occupy more books than (true) works of great literature.¹¹ Newton benefited from the developments of mathematics and physics before him (as he modestly put it, “if I have seen farther than others it is because I have stood on the shoulders of giants”), but - and this is just an intuition on my part - the Three Laws of Motion, or the theory of evolution via random mutation and natural selection, or Daltonian chemistry are accessible via more routes than Bach’s Well Tempered Clavier.

Richard Feynman, I think, put it best when he said:

If, in some cataclysm, all of scientific knowledge were to be destroyed, and only one sentence passed on to the next generations of creatures, what statement would contain the most information in the fewest words? I believe it is the atomic hypothesis (or the atomic fact, or whatever you wish to call it) that all things are made of atoms—little particles that move around in perpetual motion, attracting each other when they are a little distance apart, but repelling upon being squeezed into one another. In that one sentence, you will see, there is an enormous amount of information about the world, if just a little imagination and thinking are applied.

-Richard Feynman, Feynman Lectures on Physics, Volume 1, Lecture 1: Atoms in Motion

Think of just how many ways there are to encode that insight. An enemy of knowledge could more easily deprive mankind of The Art of the Fugue than the theory of atomism.

Another way to approach the concept of uninvention is to consider how a technology might be recovered if a cultural memory of it existed, compared to a work of art. Picture a group of philosophers in a post-apocalyptic symposium - natural science having reverted to philosophy with the loss of the scientific method - drinking heavily, discussing the technologies of the Ancients. One of them mentions a work on history that talks about a special metal that, when brought near to samples of itself while immersed in water, generates. Those men themselves are not going to recover nuclear fission in their lifetime, but now they know that it is a secret rather than a mystery - something that was once done and could be done again. If something can be demonstrated, and is remembered, then the path to discovering it again is open, even if it may take a very long time. This would meet criteria (1) for ‘accessibility’ above, being known to be possible.

If the details had been forgotten, if all that remained were films of nuclear weapons tests without additional info (bombs work by imploding specific isotopes of uranium or plutonium, for example), then the subject would be a pure mystery. Philosophers would speculate about how such effects could be achieved, but not come closer to the truth unless they had the root idea of atomism.

Alternatives

It seems like uninventing technologies at the global level is next to impossible without uninventing a lot of other technologies at the same time (to say nothing of all the people who would die in the resulting famines, plagues, and wars over the limited stockpiles of remaining resources). You can ban a technology in a specific locale if you have dictatorial control, but even then borders can be porous. The leaders of North Korea can prevent their people from making their own smartphones, but they cannot stop determined people from smuggling them over the border. Even then, it’s trying to push back against steady, persistent pressure: like a dam, it takes only one crack large enough to flood the valley of ignorance you’ve tried to cultivate.

There is one other option, not as drastic as civilizational collapse: invent a superior replacement. To return to Charles Babbage, imagine that after giving up on a 19th century Facebook delivered through telegraph lines, mediated by his Difference Engine,¹² he takes inspiration from his irritable disposition:

Damn it, Ada! These horses are driving me insane - defecating and urinating all over our streets. And people call themselves civilized, living among such filth. There must be an alternative…

What we need to do is provide an alternative means of transportation for people and goods. Perhaps we could join a steam engine to a cart, using the motive power from the burning of coal to drive the wheels…

It wasn't Babbage who did it, but the internal combustion engine and the automobile are the reason our city streets aren't drowning in horse urine. Problem solved! Except for fine particulate air pollution, carbon dioxide emissions, traffic accidents, traffic jams, road wear and tear, and so on and so forth, but the horse urine problem was fixed.

We can hope that the increasing adoption of electrical vehicles will free us from the internal combustion engines downsides, and bring less downsides of its own. A replacement technology, one notably superior to its predecessor, can practically uninvent a former one. The techniques for shoeing horses and making saddles are not lost, but they're much less common than they once were. If horses were to go extinct, or human beings were to decide that it was no longer ethical to ride them, the techniques would vanish from living knowledge.

If your issue was with the technology itself rather than the use it is put to, this is a good solution. It might also make you very rich. Then again, I don't like to think about what a ‘superior' replacement for thermonuclear weapons would be (maybe antimatter?).

Preventing the Future

Technology is a sticky thing - held in existence by human desire, curiosity, redundant encoding in the Library of Babel, as well as economic and social factors. So much for uninventing the already existent without wholesale replacement.

But what about the future? Could you prevent a future technology from being invented?

Here is a good time to stop and think about the specific terms we are using. There are lots of hypothetical technologies, from nanoscale assemblers and space elevators to faster than light travel and Artificial General Intelligence which can be conceptualized, but the path to which are not known. We can know that a space elevator needs to be built out of some kind of solid material rather than a liquid or gas, and even what the tensile strength of the material would need to be, but how to manufacture enough of a substance with the requisite strength is unknown.

Stopping people from dreaming up new technologies is probably futile, not without suppressing free speech (though you could corrupt it sufficiently to achieve the same effect; see below). But you can make the path to them even harder to achieve.

• Promote Indefinite Optimism - Peter Thiel, in his Zero to One: Notes on Startups Or, How to Invent the Future, distinguishes between Definite and Indefinite Optimism. Definite Optimism is the meta-belief that there are secrets in nature and about human beings that can be found, and that diligent searching can uncover them. Indefinite Optimism, by contrast, is an over-learning from the history of evolution and moments of scientific serendipity like Alexander Fleming’s discovery of penicillin - you can’t plan to make discoveries, so you keep your options open, throw darts at the board, and hope for the best.
  
  Plainly, Definite Optimism is going to be more productive of risk-taking innovation than Indefinite Optimism. The Indefinite Optimist, whether a person or a society, is fearful of putting too many resources into a route that may not be successful. Playing it safe is fine for investigating the borders of current knowledge, but not for making the leap to secret knowledge, those domains not directly reachable from where we are today.
  
  To promote indefinite optimism, you can build or attain a platform to disparage the ‘irresponsible’ waste of long shot research projects and promote the ideas of discovery via analogy with evolution: not putting too much into any one project, making sure students maintain maximum optionality (more consultants, less startup founders and dedicated researchers). You can achieve the same effect by directing as many bright young people as possible into aimless data mining, digging for real correlations in ever bigger haystacks.¹³
  

• Add More People to the Search - Is the search for secrets of nature, for new discoveries and technologies akin to construction and manufacturing, or is it more like computer programming? That may seem an odd distinction, but consider that programming is making new technology: the first accounting program was an application of computers that had not been made before, as was the first word processor and the first graphical user interface. A common tool was used to shape many different useful things.
  
  The naive project manager estimates that, all other things being equal, adding more people (more man-hours) to a task will lead to it being accomplished sooner. This holds true for many tasks in the physical world. With more people to man the shovels, the hole will get dug faster. With more people to work on the assembly line in more shifts, the widgets will get produced faster. The limiting cases are the # of shovels, and the # of positions in the factory + raw materials to make widgets. Surely if we want to get more science and technology, we need to just add more people (get your kids started on those STEM careers, everyone!).
  
  In programming, and I think likely in science and technology development, usually leads to failure, or at least to very long delays. Frederick Brooks’ classic work The Mythical Man-Month details how the optimistic “add more people” plan breaks down when it comes to software engineering.¹⁴ Even back in 1970, it was apparent that small teams of programmers made more progress more quickly than large ones.
  
  First, Brooks writes, there is the need to get people up to speed on the task at hand. Those who are doing the work need to take time out to teach what they know to those who have just been added to the project. This takes even more time if the knowledge is highly specialized. So whatever progress was being made is reduced during the training time. For the programmers reading this, think of how much time documenting your code takes compared to writing it (and how much less fun the former is compared to the latter, contributing to your urge to delay getting on with it).
  
  The problem gets worse the more people are added in subsequent rounds of hiring. Because the most experienced are busy, the newly hired are often tasked to train the most recently hired, with all the risks of miscommunication that brings - or worse, the inculcation of false beliefs that lead to costly, difficult to trace, and expensive to correct mistakes.
  
  If there is a shortage of programmers, electrical engineers or particle physicists, and if higher salary offers and relaxed immigration quotas cannot get more, then schools need to train them. Bright young people need to be persuaded to enter the profession; they need to acquire the education needed to make a contribution to the field. Along the way, many will be lost to attrition: some will drop out of school, others will choose other majors, still others will graduate with the degree but not practice it (compare the doctor who gets a medical degree and her license, but makes her living advising pharmaceutical companies on how to run clinical trials). Some will find employment elsewhere than on the research project. Still others will die or become disabled along the way.
  
  From this observation follows a general rule: to retard progress in ideas and technology, add more bodies to the search. If you have sufficient capital, set up grants with sufficient incentives to encourage hiring whether or not the extra bodies are of benefit to the laboratory or startup’s supposed goal.
  

• Distract From the Search With Tertiary Goals - The uneducated public assumes scientists and other researchers are in the business of making discoveries. This is not so; they are in fact in the business of publishing papers and attracting grant money to their institutions. Researchers do make discoveries, it is true, and there are many dedicated, conscientious researchers.¹⁵ But whether you are a careerist or a scholar, you have to jump through the same hoops. Worse, the selection pressure may favour those who are better at giving off the signals of high rates of publication and skills at winning grant money over those who are the most brilliant at research.

  
  (Anecdotally, I’m not sure Einstein would have had his annus mirabilis in 1905 if he had to teach undergraduates, apply for grant money, and keep up a steady stream of publications to keep his academic post. He might have just devoted his whole career to further elaborations on Brownian motion).
  
  If you have access to a lot of capital, one of the most productive ways to dilute and misdirect the pursuit of a technology is to offer grant money conditional upon filling out complicated, time-consuming applications, with intensive monitoring and plausibly related but still off-track performance goals and milestones.
  

• Prevent Hot & Cold Wars - One of the most effective, though destructive, ways for a society (a wealthy, advanced one anyway) to clear up the three above issues is through warfare, particularly Total War, where a nation’s entire productive capacity is employed in the war effort. Indefinite optimism is suppressed in favour of definite planning and goal-setting, manpower becomes scarce due to conscription and war production so teams have to make do with the people they have, and tertiary goals fall away in pursuit of victory.
  
  So far this has been a useful strategy. It gave us moderns sonar, radar, portable radios, night vision, duct tape, space exploration and satellite navigation, nuclear energy, jet airliners, digital photography, and the Internet. Arguably, today, it is one of the driving force behind many artificial intelligence advances, as competing countries seeking for am edge in cyber warfare fund university AI research.
  
  If you had the power to effectively promote global peace and comity, then you could plausibly retard scientific and technological progress through peace.¹⁶
  

• Corrupt Language - In Orwell’s 1984, the Party entrenches its dominance of Oceania through the gradual reduction of the language. Each edition of the official dictionary has fewer words, reducing the range of expression. If it is harder to talk about an area of inquiry, then its harder to make concepts that push the boundaries, and harder to make progress in understanding and invention.¹⁷ Promoting speech codes that limit freedom of expression under the guise of protecting vulnerable groups, or flat out prevent the publication of research about certain topics for the same reason. Either being in administration, or having the capacity to manipulate administrators, is a powerful position for the uninventor to occupy.

We’re Not Actually Here to Uninvent Technologies

I hope you’re now saying “that was fun!” but I think you may also be saying “but what now? Why did I read several thousand words on suppressing technologies?” As you could probably guess, the above is not a serious manual on how to suppress technologies whether established or forthcoming. Rather, it’s an exercise in inverted thinking - trying to think about how to improve the rate of technological and scientific progress by thinking first about how to retard it.

Make no mistake: we need new technologies. As Peter Thiel points out in Zero to One: Notes on Startups Or, How to Build the Future, progress comes in two varieties:

1. 0 to 1

2. 1 to N

Going from 1 to N is just doing more of what we are already doing. People need electricity? Build more coal-fired power plants. People need more transportation? Build more gasoline-powered vehicles. People want more seafood in their diet? Send more boats into deeper water with larger nets. The problem is 1 to N by itself is not sustainable on any finite planet.

Going from 0 to 1, on the other hand, is the key to both sustainability and real material progress.

The transition from wood to coal burning in England just before the Industrial Revolution, for instance, was not great from our modern standpoint dealing with the accelerating effects of climate change. For those living at the time, however, the transition spared the English countrysides remaining forests, and allowed many to grow back over the ensuing centuries. Though it smelled much less pleasant, coal also provided more megajoules of energy per kilogram than wood. It was a 0 to 1 change.

The old way of doing things, cut more trees to make more heat, was 1 to N. Switching to coal was (in the beginning) 0 to 1. Building more nuclear power plants to replace the coal fired ones is, for us, 1 to N, because nuclear power is, by now, an old technology - not all 1 to N’s are equivalent, however, as replacing all of our conventional, fossil fuel power with nuclear power would be a major win on the greenhouse gas emissions front. New nuclear technology, whether Generation Four power plants or (someday, somewhere far over the rainbow) fusion power, would be going from 0 to 1.

Nick Bostrom provides an apt metaphor for our predicament:

When a rocket stands on the launch pad, it is in a fairly sustainable state. It could remain in its current position for a long time, although it would eventually be destroyed by wind and weather. Another sustainable place for the rocket is in space, where it can travel weightless for a very long time. But when the rocket is in midair, it is in an unsustainable, transitory state: Its engines are blazing and it will soon run out of fuel. Returning the rocket to a sustainable state is desirable, but this does not mean that any way to render its state more sustainable is desirable. For example, reducing its energy consumption so that it just barely manages to hold stationary might make its state more sustainable in the sense that it can remain in one place for longer; however, when its fuel runs out the rocket will crash to the ground. The best policy for a rocket in midair is, rather, to maintain enough thrust to escape Earth's gravitational field: a strategy that involves entering a less sustainable state (consuming fuel faster) in order to later achieve the most desirable sustainable state. That is, instead of seeking to approximate a sustainable state, it should pursue a sustainable trajectory.

The present human condition is likewise a transitional state. Like the rocket in our analogy, humanity needs to pursue a sustainable trajectory, one that will minimize the risk of existential catastrophe. But unlike the problem of determining the optimum rate of fuel consumption in a rocket, the problem of how to minimize existential risk has no known solution.

-Nick Bostrom, Existential Risk Prevention as Global Priority

To torture Bostrom's analogy: 1 to N is adjusting the pumps feeding fuel to the engine, increasing the rate of consumption (fossil fuels burned, percentage of atmospheric carbon sink used up) without increasing efficiency. We know there isn't enough fuel, in the current engine, to get us to that stable orbit. Going from 0 to 1, on the other hand, is modifying and increasing the efficiency of the engine. It's risky and you could blow up the rocket (just look at the literature on geoengineering and AGI safety) but without greater efficiency in using the fuel we're not going to make it to a stable orbit.

We need to go from 0 to 1. And if we're not in a position to pursue the advances ourselves, we need to get the obstacles out of the way of the inventors. The first way to do that is to stop the (functional if not deliberate) uninventing.¹⁸

1. Why not include anthropogenic climate change on the list? I’m glad you asked. It is because, of all but the most catastrophic IPCC scenarios, the bad consequences unfold over decades, with increasing spikes of severe weather. That is a long enough time frame for much of humanity - though sadly probably not all of it - to adapt to the new conditions, using the technology we have. Barrier walls to stop rising sea levels, solar radiation geoengineering, and rapid adoption of low-carbon power sources are possible over decades. I am not sanguine about this, it’s just that the truly civilization wrecking events need to be fast to be effective.

That’s why my primary worry about a changing climate is not severe weather, droughts, or even famines, it is what these conditions will lead nation states to do for self-protection. Climate change will stress the world’s states in ways they never have been before, though not all to the same degree at the same time. It likely will push some states into open conflict with neighbours, and if those states have nuclear weapons the chance of a catastrophic miscalculation or deliberate atrocity increases. Climate change is not the worst scenario - climate change with even a limited nuclear war (antinuclear activists scoff at this notion, but it is more likely than an all out nuclear exchange between two superpowers) could push us too far past the point of stability too fast to adapt to a changing climate. Not to mention the added problems of climactic instability caused by the exchange (IF nuclear winter is an actual possibility, a contentious point).

2. The part that goes “boom,” and is separable from the delivery vehicle - gravity bomb, ICBM, SLBM, or U-Haul truck.

These repairs are difficult, and not just for the obvious “nuclear weapons are complex pieces of technology” reason: the original weapons were designed like Apple products - use for a few years, then throw away when you get a better model. As such, they were not designed with such conveniences as ports that let you easily access their inner workings or components that are easy to take out and replace if they break. This was not a problem until the United States put a moratorium on both nuclear testing and manufacturing new nuclear weapons. Before that, if you wanted to know whether a particular type of weapon was still functional, you dug it out of storage, stuck it down a deep hole under Nevada, and set it off. If Nevada gained one more crater, you gained more statistical certainty that the rest of the devices were still sound. Without testing, you need to do delicate, complex, and expensive component testing and computer modelling to be able to say “yeah, probably…”

So until total nuclear disarmament happens, nuclear testing is resumed, or the United States begins designing and building new nuclear weapons, they’ve got to make the ones they have last as long as possible.

3. Interestingly, the plastic of the case was the very hardest part. In talking the idea over with an Imperial College of London professor, a specialist in mining technologies, Thwaites learned that the plastic was the part he would not be able to make from scratch himself. Despite being highly ubiquitous, our supposedly commonplace plastics are exceptionally hard to make, not the kind of thing even an accomplished chemist could do on a lab bench in their shed. If you’re curious, the British Plastics Federation has a high level overview of the manufacturing process for PVC.

4. Specifically, it is likely to be a packing material inside the radiation case between the primary fission bomb and the secondary store of lithium-deuteride, which fuels the fusion stage of the explosion. Arms control experts believe the aerogel material explodes under intense X-ray bombardment, ensuring that maximum force is applied to the secondary before the detonation fractures the radiation case and blows the components apart.

5. Borges does not detail the proof within the story. A shame.

6. To get you started, assume each page of the book can contain text - a fully blank page is just a page with spaces. That yields 410 pages per book. Each page is 40 lines long, and each line has places for 80 characters. That’s 3200 character spaces per page x 410 pages = 1,312,000 character positions per page. Each of those characters positions can have 25 possible orthographic characters. Repetitions are allowed. Keep in mind that while books cannot repeat, individual pages may, both within the same book and between books.

And I think we can stop there and agree that the Library is inexhaustibly vast. No one is finding a work of literature, or even anything comprehensible, by random chance. Then just imagine its size, and the number of books, if it used ASCII. Or even worse UTF-8.

7. It’s a neat exercise to think of all the things that are in the Library of Babel. All the greatest books ever written, and the greatest books not yet written, and the greatest books that will never be written. Every genetic sequence of every living creature from the dawn of life on Earth to the final extinction, as well as genetic sequences that will never be realized, and alien genetic sequences. Pseudo-code descriptions of every program it is possible to write. A complete map of all of the interconnections of a human brain (described, tediously, in text, across hundreds of volumes). The best possible translation of each work of literature into every language. The solutions to the Millennium Problems.

As for true things that are not in the Library, those would be truths that are not apprehended by reason or language. Though presumably the Library contains books that would point the way towards them for a dedicated seeker.

8. Computer scientists, if I’m wrong, let me know in the comments!

9. I’m not saying here that a work of art can’t be reduced or refined - this is what authors do when they edit their works, after all - but that only a limited range of variants have the aesthetic resonance of the canonical edition, the one from which the others are corruptions. Our world may not even (and here is a depressing thought) possess the ‘real’, canonical versions of our greatest works of literature and music - the creators got close, but not all the way there. This is certainly how many authors feel about their published works, that they didn’t quite succeed in what they set out to do.

Often, as a writer, I feel I am groping around trying to capture a canonical version of a story I am only able to feel my way towards, and must experiment to put in the right way.

Tom Stoppard’s play Arcadia makes this connection between poetry and physics. As a poet has an intuition of order that they work their way towards, swapping out words to make the poem ‘sing,’ so a mathematician follows an intuition to create the formula that captures the phenomena they are studying, swapping out constants, variables, and operations till he arrives at the truth that captures his insight.

10. I am aware that Shakespeare’s works have come down to us in separate Folios, with textual variations between them. That means there is a King Lear as written by Shakespeare, but our access to it is mitigated by potentially faulty copies.

11. I regard Truth, Beauty, and Goodness to be synonyms for Being. So an aesthetic truth and a scientific truth belong to the set of all truths. Some are just more redundantly encoded in our world’s possibilities than others.

12. Example is from a previous essay, Twice Read Books: Peter Thiel's Zero to One” or, On Secrets and Mysteries

13. Not that data mining can’t achieve great things when done properly, its just that it is an indefinite optimism activity, not the definite optimism of seeking a (gamma) secret based on a truth you believe in that few other people agree with you on. See the link in footnote 12 for more on the distinctions between grades of secrets, and their distinctions from mysteries.

14. A “man-month” is a unit of estimating work, similar to man-hours. It was commonly used, when Brooks wrote the first edition in the 1970s, to estimate the potential productivity of an industrial worker.

15. A common theme I am noticing in my thinking and writing: “I think poorly of X as a profession, but well of S as an individual X.” e.g. I think poorly of teaching as a profession, but well of individual teachers, poorly of politicians as a class but well of individual politicians, and so forth.

It’s not clear to me if or even how the inverse could be true: “All the individual S’s are pieces of shit, but the S’s as a professional class are saints.”

16. This one I am much less sure about than promoting indefinite optimism, adding more people to the search, and setting up distracting tertiary goals. Plausibly, in an era of world peace and stability, the money currently lavished on the arms industry would be greatly reduced and redirected into more productive areas like medicine, automation, and asteroid mining.

17. The London prison psychiatrist and essayist Theodore Dalrymple eloquently describes the plight of those with limited vocabularies:

“With a very limited vocabulary, it is impossible to make, or at least to express, important distinctions and to examine any question with conceptual care. My patients often had no words to describe what they were feeling, except in the crudest possible way, with expostulations, exclamations, and physical displays of emotion. Often, by guesswork and my experience of other patients, I could put things into words for them, words that they grasped at eagerly. Everything was on the tip of their tongue, rarely or never reaching the stage of expression out loud. They struggled even to describe in a consecutive and logical fashion what had happened to them, at least without a great deal of prompting. Complex narrative and most abstractions were closed to them.”

-Theodore Dalrymple, “The Gift of Language” (City Journal, August 2006)

Available online here

18. Though please, please, please don't start wars as a means to accelerate technological progress.