Materials and Manufacturing Underpin Civilization

My talk from Progress Conference 2024

Why is it so important for Speculative Technologies to focus on materials and manufacturing? Here’s the essay version of my talk from the first Progress Conference (video version here) where I try to lay out the thesis behind our focus more fully.

To be maximally provocative: the materials in the device you’re reading this on and the processes that create and assemble them are more important to progress than the smallpox vaccine.

When you think of technologies that “help people” or save the world” what comes to mind? For most people, it’s things like life-saving medicines, carbon sequestration, or answers to existential risks. But we can’t do any of those things without new technologies in materials (“stuff”) and manufacturing (“ways to turn stuff into things”).

Consider stainless steel. In addition to being useful for easy-to-clean utensils, cookware, and sinks, it’s what enables the syringes that deliver your vaccines, the homes for the e-coli that create insulin, and many medical implants. Stainless steel wasn’t widely available before 1915, after a long and winding process of science, invention, diffusion, and scaling.

Further back in time, the power loom made it so that instead of freezing to death, it’s more common for people have too many clothes; not to mention clothes that are softer and better designed than people could imagine in the past. Iron enabled us to cultivate so many more soils and make things like nails that we don't even think of as “something we once didn’t have the capability to make.” Bronze sickles enabled humanity to drastically increase the calories we had access to. The importance of materials and manufacturing predates written history: stone and our ability to manufacture tools out of it enabled us to spread across the world. There’s a reason we literally name eras after the materials.

This kind of progress is not inevitable. In the same way that the eradication of smallpox or near universal literacy required great undertakings and agentic people, so too does the creation of new paradigms in materials and manufacturing. If you want more progress, whether that looks like Dyson spheres, happy healthy people playing basketball with their grandkids, or some other form of flourishing you need to care a lot more about boring materials and grimy manufacturing.

Why Materials and Manufacturing are important

In a nutshell, the properties of the materials we can harness, along with our ability to make them at scale and turn them into useful artifacts (ie. manufacturing) circumscribe what humanity is capable of. Creating and deploying new materials is the bedrock of all other progress. What do we need to unlock hypersonic flight and cheaper space travel? Better materials. What will we need for seamless fusion reactor maintenance or solar panels you can unroll like sod? Better materials. Everything from super capacitors that could charge an electric car in seconds to space elevators are a dream until we figure out new materials and manufacturing. Brain-computer interfaces need electrode materials that don’t foul. The cost of new kinds of medicines is often limited by our bioreactor technology. The list can go on ad infinitum but the point is that materials and manufacturing feed into literally everything else we associate with abundance.

While many people think of them as different domains, I talk about materials and manufacturing in the same breath because they are two sides of the same coin: a snake eating its tail. Steel was too expensive for anything but mission-critical items like armor or swords until Henry Bessemer created a new manufacturing process that made steel cheap enough that it became one of the most used materials in the world; in 2023 we produced 1.9 billion tons of the stuff. At the same time, biomanufacturing would be impossible without stainless steel and plastic’s anti-corrosive properties.

Materials and manufacturing are increasingly abstracted from most people’s lives. We think about the energy grid every time we pay an electric bill or the power goes out; most people interact with some form of transportation every day, even if it is just hearing a car; we all carry around internet-connected electric brains; and even though farming feels the most analogous to manufacturing people think much more about how their food is made and where it comes from (organic! Grass fed! Local! No GMOs!) than about how their widgets are manufactured. This abstraction has some potential downside: it has perhaps led to a situation where smart people no longer do base-layer work and manufacturing is important and often ignored.

Let’s de-abstract materials and manufacturing some stories:

Carbon fiber is an incredible material that is effectively made out of soot but has a better strength/weight ratio than steel. Its story starts roughly in 1958 when Roger Bacon (the less famous one) created carbon fibers; but they were only 20% carbon — you need near 100% to be full strength. From the beginning, researchers realized that these fibers could potentially replace aluminum in aerospace applications. In 1960 researchers figured out a manufacturing process that could get the fibers to 99% carbon but it wasn’t until 1968 that they were incorporated into products: fan blades that failed and almost drove Rolls-Royce out of business. In parallel to this Japanese researchers figured out how to make carbon fiber yarns instead of individual fibers which were used in several useful but low margin products. In 1988 Jacob Lahikani figured out how to make composites with carbon fiber and epoxy, which mitigated many of the downsides of raw carbon fibers but was incredibly labor intensive to manufacture. It took another 20 years to figure out how to make carbon fiber composites at large enough scale that engineers could finally realize the dream of carbon-fiber airplanes — the Boeing 787.

Stainless steel’s malleability and corrosion-resistance unlocks everything from syringes to bioreactors to easy-to-maintain silverware. Its story arguably starts in 1798 when Luis Vauquelin isolated chromium as an element. In the early 1800s Michael Faraday and others observed that chromium-iron alloys resist oxidation. Over the following decades, Krupp steel and other manufacturers attempted to incorporate that observation and make corrosion-resistant steel. However, they couldn’t add chromium to steel without also introducing additional carbon, which mitigated most of the chromium’s corrosion-resistant effects. (Small changes in composition can make a big difference!) It took another several decades for two discoveries to unlock stainless steel in the 1890s: Adolphe Carnot (nephew of Carnot Cycle Carnot) discovered that reducing the amount of carbon in steel increased its corrosion resistance and Hans Goldschmidt invented a process to make chromium without additional carbon. It then took another decade for manufacturers to incorporate these discoveries into processes that could make high-quality stainless steel at scale.

There are several themes that run through these stories:

• Progress in materials and manufacturing happens through cycles of both invention and discovery: trying to create a useful thing based on a discovery about the nature of the universe raises new problems that can only be solved by discovering more things about the universe (which often requires inventing other things). These cycles fly in the face of how we often talk about technology creation as “basic research which leads to applied research which gets transitioned into a company that does development.”

• Progress in materials and manufacturing requires contributions from many different people and organizations, most of whom don’t capture the eventual value.

• In part because of the first two points and in part because it requires a ton of trial-and-error in the physical world, fully operationalizing a new material takes a long time; 40 years passed between Rolls-Royce’s disastrous attempt to create carbon fiber airplane propellers and the successful flight of a carbon-fiber 787; it took 75 years from Krupp’s attempts to create corrosion resistant steel in the 1850s to the patent on stainless steel in 1915. While we can certainly hope to speed things up, the baseline is that materials and manufacturing innovation operates on timescales that are very different from what we’ve come to expect from software.

To de-abstract manufacturing, I want to show you an image of not even all the components that go into a refrigerator:

It’s easy to abstract away how things are made into “oh yeah someone just makes a refrigerator” but there are dozens of components and multiple processes that go into creating each component, with different firms and supply chains specializing in each process. Here is a list of all the different manufacturing processes at a recent trade show to give you a sense of everything behind everyday items.

We’re going to need a lot of new materials and manufacturing processes to build the future! If we briefly take the “future if…” meme seriously, we can peel back the flying cars and robot dogs to see that under the surface, it’s new materials and manufacturing processes that unlock this world (or any abundant future).

And yet, if we look around us, most of the materials we see are the same as in the 70s. Most of the things around us were manufactured in the same way too. Despite their importance, the rate at which we’ve upgraded our materials and manufacturing processes has declined. This engine of progress has stalled out.

Like many stalled engines, this isn’t due to any single failure point, but a slow accumulation of frictions. Rhetorically, I should be telling a nice clean story here, but I want you to come away with the realization that the story is not clean and there is no one weird trick.

Instead of unpacking a list of issues, let me tell you a few heavily redacted stories that are meant mostly to give a sense of the vibes. (If you’re interested in a more detailed treatment of the current frictions, you can check out this piece I wrote in the Works in Progress Magazine)

1. There was a company that was trying to make a “McGuffin” that was something people have dreamed of for decades. They legitimately figured out how to do it — I saw it work myself — but it involved very carefully coupling one widget to another by hand. Not only could it only be done by hand but there was one guy who could do it and he wasn’t able to teach anybody else how to do it. The McGuffin needed to be made at scale, so a manufacturing process where one guy makes the thing by hand was totally untenable. However, the company was under so much pressure that they didn’t have the space to do the research to figure out a process to manufacture the thing at scale — instead they went down a much more conservative technological route that made them no different than anybody else.

2. When I was helping people start companies in Singapore, I would often run into a situation where an ambitious researcher would tell me an idea. In response I would say “ok, first let me put on my investor hat. This is a terrible idea for a startup — nobody will invest in it, it will take five years of research before you know if it will work, and even if it does there isn’t really a product there” then I would pause and say “ok, now I have my Ben Reinhardt hat on — this is amazing and the world needs it, a startup just isn’t the right way for this to happen but neither are other institutions like academia or big companies.”

The big picture is that timescales, capital costs, and level of integration with other technologies makes materials and manufacturing are particularly susceptible to the valley of death and friction in the system has made it so that valley is wider and deeper than in the past.

The closest thing to ‘one weird trick’ we have is to ask “what did environments that were especially good at unlocking new materials and manufacturing have in common?” And “how can we build new ones?”

It’s easy to cargo-cult Bell Labs, but they legitimately did create amazing things. A lesser-known story in the creation of the transistor was that Walter Brattain was completely stuck because he had no idea how to get good contact between a metal and germanium. He happened to be complaining to a machinist, Clyde Hutchinson, who had seen problems like that before and had incredibly deft hands that enabled him to create the contacts and unlock the first transistor.

During the creation of the Haber-Bosch process for fixing nitrogen (and therefore feeding the world), Franz Haber did a proof of concept, but it was useless unless it could create millions of tons. The prototype used cobalt as a catalyst, which, at scale, would have required more than the entire world’s supply of cobalt. Haber’s team needed to figure out a scalable catalyst, the thermodynamics at scale, and solutions to problems that only presented themselves when the process was running at full scale and for long durations: over time, hydrogen would embrittlement the metal walls of the high-pressure reactors, causing them to explode.

Some common traits among these enabling environments that are now rare:

• They have research, prototyping and manufacturing expertise all in the same room, all with aligned incentives.

• Individual mavericks can get stuff out the door quickly.

• The ability to work on things over long timescales.

(As a brief plug because it’s my job) That’s the kind of environment we’re trying to build at Speculative Technologies: a home for research misfits focused on building useful technologies that aren’t good fits for academia or startups.

We can’t do it alone. Unlocking the future of materials and manufacturing is going to require building many new kinds of institutions and clearing out old ones — from new ways of training people to new financing structures and reducing regulatory friction.

It’s easy to feel like progress in materials and manufacturing is inevitable, that if we are just optimistic, markets will work their magic. Progress is not inevitable. Consider the Antikythera mechanism: an orrery from 2nd century BC Greece discovered on a shipwreck in 1901. It’s an incredibly complex geared device but completely handmade. If the Greeks had figured out how to manufacture gears and mechanisms at scale, we might have had an Industrial Revolution two millennia early. There are so many technologies that are potentially the Antikythera mechanism of today.

Even if I haven’t managed to convince you that it has become harder to get materials and manufacturing technologies into the world, the Antikythera mechanism suggests strongly that we can and should do a better job unlocking the technologies that underpin civilization.

As Patrick Collison mentioned during the progress conference, Progress Studies is in large part a vibe. So I want to leave you with the vibes that I’m going for with all of this: materials and manufacturing unlocking new frontiers that enable more progress.

Every time you get a vaccine, thank stainless steel. When you marvel at a new AI model, thank the silicon and copper that enable it. When someone talks about unlocking fusion power or uploading consciousness or extending lifespans, ask them about the materials that will make it possible.

We started with a provocative claim: materials matter more than the smallpox vaccine. But that understates the case - without new materials and manufacturing, we won't get the next vaccines, the next computers, or the next revolutions in human capability.

Any sufficiently advanced technology is indistinguishable from magic, but there's nothing magical about it - just atoms arranged with exquisite precision through processes we've spent centuries perfecting. The future doesn't arrive just by focusing on charismatic technologies with obvious connections to people’s wellbeing or a bottom line. It arrives through mastery over matter itself. That's what we're working on at Speculative Technologies. The most important work you've never thought about.

At Speculative Technologies, we're not just working on materials - we're working on unlocking the future itself.

Want to help? You know where to find us.

As a shameless end-of-year plug: Speculative Technologies is a 501(c)(3) supported by people who want to unlock the future. We want to be the best way to turn resources into public goods that open new branches of the tech tree. You can donate here or email info@spec.tech for larger donations.