Editor’s note: I wrote this to provide context to my summary:

I am a researcher and engineer investigating core problems in artificial intelligence, cognition, and sensing.
I target applications in computer vision, robotics, and natural language.
I work on frontier technologies, creating new IP and original research.
I transfer tech from lab to product.

---

I’m not interested in working on your latest social media app. I don’t want to add a new button to your website. I don’t want to create the next TikTok. These are tech-enabled businesses.

I work on Hard Tech. My work is to develop core science and build engineering advances. My goal is to create products that leverage technological advantage.

Hard Tech vs. Tech-enabled

Hard Tech companies must solve significant science and engineering problems to create their products. They and their competitors are ‘rate limited’ by scientific advancements. R&D spending is significant. The amount of top talent with these highly specialized skills is limited. Barriers to entry are high, because of talent acquisition, capital requirements, and long cycle times to fill the R&D pipeline, which is essential to creating new products. As a result, oligopolies are common. The archetypical hard tech industries are semiconductors and nuclear power. Because of the competitive landscape, stagnation can also occur. The aerospace industry is one such example, with minimal change in consumer experience in 25 years.

In contrast, most tech-enabled businesses have a low barrier to entry. A number of tech-enabled businesses have grown rapidly – so they now face hard tech scaling issues. For example, Facebook handles 350M photo uploads a day and has 8 billion video views a day.

Hard Tech often serves as a platform for other Hard Tech and tech-enabled businesses. For example, regular reduction in CPU energy consumption resulted in better, longer lasting, more powerful, smartphones. This enabled Uber. It also enabled physicists and chemical engineers to create better material simulations, which contributed to creating the next generation of Intel CPUs. An advancement in the base platform yields dividends for all users. And it automatically comes to dominate the market – nearly nobody does serious computing on CPUs older than 4 years.

Hard Tech Entrepreneurship

New technology, when applied correctly, generates value. As an entrepreneur, you capture some of this value (via pricing) and you pass some of it on to your consumers (via savings). In this way, technological advantage creates pricing power. This is why so many of the most profitable companies in the world are technology companies.

Those who leverage advanced technologies gain an advantage over competitors, and can reinvest their savings into more advanced technology. The technology may allow them to participate in technological frontier markets, like the Internet in the 90s, or perhaps Bitcoin in the ‘10s. This creates a self-reinforcing cycle where returns accrue to those investing in advanced technology. Therefore, a sufficiently technologically advantaged new product will automatically own the market, because of the superior value it provides consumers.

I don’t want to compete over how ‘refined’ my UI is, or the quality of my product’s packaging. I don’t want to compete over who has the best viral marketing. If I have technological superiority, I don’t need to invest my limited resources in these business directions. As my competitive advantage is in technology development, I am better off anyway. The technological advantage dominates the value proposition of my product; other improvements may make only marginal contributions to the value proposition.

Violating Assumptions

Advancements in Hard Tech compound so quickly as to break assumptions. Below are several statements that would have been considered ridiculous as recently as 10 years ago. Taking and sharing a photo is free. Flying is cheaper than driving. Phone calls are free. Lighting is cheap, small, can be placed anywhere, and has near zero running costs. I never need to physically visit my bank. I have a personal assistant that gives me driving directions, accounting for traffic, instantly, for free. My car never needs maintenance. My house is vacuumed and mopped by a robot.

Hard Tech is the driver behind these changes. That’s a compelling reason to work on hard tech.

Why Hard Tech?

What assumptions do we have today that will be violated?

It may be, “I need a driver’s license.” I already drive less, because I don’t need to go to retail stores. I can order groceries online. And with the pandemic, I don’t need to go to the office.

Self-Driving cars will of course be the final nail in the coffin. They will free people’s time and money from the automobile. (Personally, I love driving). SDVs will launch a legion of Tech-enabled businesses. Think about in-car entertainment and advertising. In-car productivity. The restaurant that pays for your SDV trip from home to the restaurant. And the SDV RV, so you can sleep in peace while being driven home instead of flying. In this way, SDVs will serve as a platform for other businesses.

Working on the hard tech ideally positions you to be able to create more hard tech, or tech-enabled businesses, both based on the hard tech platform you helped create. This is because you understand the underlying technology the best. Specifically, you understand exactly how the compounding gains manifest themselves, and can therefore maximally take advantage of this free ride. This is “knowledge vertical integration.”

Conclusion

In summary, Hard Tech:

• Requires solving significant science and engineering problems.
• Creates significant pricing power.
• Creates significant barriers to entry.
• Creates value for creators and consumers both.
• Results in a platform effect with compounding improvements.
• Positions you to create more hard tech or tech-enabled businesses, because you understand the hard tech platform.

In a future post, I will go into how computer vision, robotics, and natural language are benefitting from the hard tech as a platform effect, and why that makes these fields excellent target applications for hard tech.