Manufacturing & Intellectual Dark Matter

Capital is not sufficient to re-shore manufacturing. We need knowledge & talent.

I recently came across a concept from Samo Burja called intellectual dark matter. In the physics context, the idea of ‘dark matter’ is used to explain phenomena that exist in space which are unexplainable without the presence of some large quantity of matter which we cannot see with either our eyes or our instruments. We know it’s there, but it cannot be accessed or explained directly (at least not yet).

Burja applies this concept to societal knowledge. There is a large amount of knowledge which we cannot see, though it has to exist - lest our social fabric start ripping at the seams.

3 Types of Intellectual Dark Matter

1. Trade secrets. Some knowledge is hidden because it gives businesses a competitive advantage. Biotech companies, tech firms, and Coca Cola are all keeping proprietary information under wraps for financial reasons. Most western societies believe that protecting this knowledge from public exposure serves as a good incentive to create it in the first place, so we’ve set up laws which help keep this information hidden.

2. Tacit knowledge. While some knowledge, like trade secrets, are kept hidden intentionally, other forms of knowledge are just hard to document and explain. You can see this in professional athletes who cannot accurately describe what they’re doing when teaching younger players. This is also evident in many forms of manufacturing work (which we’ll get more into later). Burja uses the example of the Bessemer steel process, which was invented as a way to make steel more efficiently. After licensing the process to other companies, Bessemer was called a fraud because these companies couldn’t get his process to work on their own - despite Bessemer’s concrete instructions. Bessemer had to set up his own company to produce steel himself using the method to prove its validity.

3. Lost knowledge. This is the scariest kind of intellectual dark matter. According to Burja, we only have the writings of just 13% of the ~2000 Greek authors known to us by name. And of these 13%, we only have partial works. For example, we don’t have the entirety of Aristotle’s book on economics. How much did the Greeks know that society had to relearn? Where would we be today if the Greeks were able to successfully pass 100% of their knowledge down to future generations?

We Know How to Preserve and Grow Knowledge in Software

Open source software and crypto do an excellent job of preserving knowledge and preventing key information from being lost or hidden. There are many successful Youtubers who help students learn the tacit aspects of software development, much of software is open source (big companies even spend millions on open source development), and there are efforts - particularly within crypto - to prevent knowledge from being lost (check out Arweave and IPFS for specific examples of this). 

But our future will depend on more than software. We need to build in the physical world too. Software alone cannot solve our manufacturing problems. I love software - but we need to think about how we’re going to build stuff.

Globalization & Fractured Manufacturing Knowledge

The West outsourced large swaths of its manufacturing base to China and other low cost labor environments over the last 30+ years. The globalization narrative ran strong for decades. A peaceful, relatively stable period of global relations led the world down the path of just-in-time manufacturing and hyper-efficiency.

But COVID-19, the Russia/Ukraine war, and tensions between the US and China are putting immense stress on the system. And there is little redundancy built in, because Wall Street views redundancy as inefficiency. And inefficiency doesn’t lead to optimal quarterly returns. The average American executive holds their position for 5 years, and they’re largely compensated based on the company’s stock price. Outsourcing and removing redundancies is what they’re incentivized to do.

The American elite did remarkably well as a result of globalization. Asset prices are way up (even taking into consideration recent selloffs in mid 2022), but adjusted wage growth is roughly constant. The gap between rich and poor is wide. Profits stayed in the US, but some types of jobs did not.

As someone who grew up in the outskirts of Chicago and spent 7 years living in Ohio, I’ve seen hundreds of rusted out, abandoned factory buildings. I’ve seen once thriving factories closed down. I’ve seen countless towns that have been beat down by this dynamic.

Some of this is due to automation, to be sure. But a lot of it is due to globalization. Globalization isn’t inherently bad. It leads to many things being produced at lower prices, and in higher living standards for the countries exporting goods. The issue with this trend is that it has led the US to not just lose manufacturing jobs, but knowledge of how to build in the physical world at scale. As more factories close, more local manufacturing knowledge is lost, leading to more factories closing, and more knowledge lost. It’s not lost in the sense that there isn’t some person, somewhere in the world with the knowledge. It just might not be available locally.

There are signs that the trend toward globalization is reversing. When supply chains start breaking down due to war, pandemic, or future problems due to climate change, pure free trade is inhibited. And if civilization becomes more divided, that will pose problems for the US and the West more broadly because we’ll need to relearn how to build.

I’m not here to advocate for tariffs or anything of the sort right now, but I am here to shine light on the fact that producing things is about more than capital allocation - it’s about knowledge. It seems to me that many Americans think that the US could produce most things domestically, but we that we just choose not to because it’s cheaper to do it elsewhere.

This cuts at the heart of a misconception that I think many Americans have about how this works - a misconception that I had as well.

We believe that we can just flip a switch and start producing things in America again. Sure, it might be more expensive, but we could do it, right?

Right?

Knowledge Is More Important Than Capital

Andy Grove, the great head of Intel, argued that it was of vital importance to keep manufacturing and design close to one another.  It enabled those in design to come up with things that were more likely to be produced at scale, and for manufacturing engineers to have a better understanding of how they could tweak production lines to enable greater degrees of design freedom. Losing manufacturing would mean breaking this loop, and ultimately inhibiting innovation.

Here’s an excerpt from an NYT article on Grove shortly after his death in 2016:

“Mr. Grove contrasted the start-up phase of a business, when uses for new technologies are identified, with the scale-up phase, when technology goes from prototype to mass production. Both are important. But only scale-up is an engine for job growth — and scale-up, in general, no longer occurs in the United States. “Without scaling,” he wrote, “we don’t just lose jobs — we lose our hold on new technologies” and “ultimately damage our capacity to innovate.”

Grove was in the semiconductor business. Today, semiconductors play a key role in geopolitics. The looming threat of a China/Taiwan conflict would mean that semiconductor exports to the West could be halted or limited, which would give the CCP a leg up in an arms race and prevent the US tech industry from making hardware advancements. And let’s not forget: semiconductors are in everything now. Cars, planes, phones, laptops, factory machines, and medical devices all need chips. 

The US recently passed the CHIPS Act as an effort to re-shore semiconductor manufacturing. Morris Chang, the founder of TSMC (the world’s leader in semiconductor fabrication), responded by saying that the US does not have the talent needed to make this process successful. TSMC is going to build more US fabs, but Chang believes that US efforts to re-shore semiconductor manufacturing will be “an exercise in futility.”

Losing much of the American industrial base meant that we lost both proprietary and tacit knowledge which was once generated locally. We traded knowledge for profits. The US manufacturing base is staring down a problem of intellectual dark matter.

And it’s not just semiconductor manufacturing that poses an intellectual dark matter problem for the US. Hadrian, a rapidly growing manufacturing startup based in Los Angeles, is tackling the problem of precision parts manufacturing. Much of their work will service the American space and defense industry. A major reason for the company’s founding is due to an insight that founder & CEO Chris Power identified after making thousands of cold calls to US plant managers.

Power identified that the precision parts supply chain for industries such as energy, space, and defense was supported by thousands of small machine shops spread throughout the country. Many of these companies were founded during the first space race, and have owners who are nearing retirement age. These shops are not advanced digitally: many of them are still using pencil and paper instead of organized CAD files - so private equity companies can’t just acquire & repurpose the IP. 

And here’s the thing: these machine shops are doing really, really advanced work. They’re building key components for SpaceX rockets and F-35 fighter jets. Running these kinds of businesses requires a deep skillset that one can’t acquire overnight. Unfortunately, due to the fact that most young Americans aren’t exactly jumping at the opportunity to pick up these family businesses, there’s a real risk that many of these companies soon wind down and shutter their doors - leaving the knowledge to die with the business, turning it into intellectual dark matter of the lost variety.

Re-Shoring Will Help Preserve & Grow Manufacturing Knowledge

Manufacturing is not sexy, but it’s necessary. For us to live in the future that we all want to live in: one that’s abundant in not just the digital, but physical world, we’re going to need it. It’s also critical for national security. A major reason for the allied victory in WWII was the productivity of the American industrial base. But you can’t just throw money at a problem like this. It requires human capital and knowledge. Developing it takes time.

We need to keep in mind that we are at risk of losing knowledge that prior generations fought very, very hard to acquire. Ideally our ability to share and preserve knowledge in software will apply to what we can do with hardware. One day we’ll live in a future of mostly automated factories that make food, housing, and consumer goods abundant. 

To do this, we’ll need to dig up intellectual dark matter in manufacturing, and prevent it from being lost.