Improving Human Systems

Hybrid Systems

About Hybrid Systems

Hybrid systems

Def. A hybrid system is a system that integrates humans and computers working towards a common goal. Hybrid systems outperform all other (existing) systems.


My thesis

I’m an investor, and it is well known that investors must have at least one thought piece. This is mine. Enjoy!

We invented society about 14,000 years ago. A little bit later, about 80 years ago, we invented computers. This has made a lot of people very angry and has been widely regarded as a bad move.

Because computers are so new, we haven’t yet had a chance to integrate them everywhere. So, as of 2023, there are lots of places where humans are stuck at their computers completing monotonous, high-precision, time-sensitive work. It’s a waste of human potential — and it’s wasteful. Humans shouldn’t be cogs in the machine. Being cogs in the machine is why we invented cogs and machines.

Often, replacing the “human computers” doing monotonous, precise work has an exponential effect: it’s not just saving one salary, it’s transforming the fundamental assumptions of the whole system.

That’s the investment thesis: I look for products, services, and companies that transform an industry by changing a human system into a hybrid system. After that, I put on my monocle and say things like “but what’s the market size???”


Deliver Hybrid Value

Hybrid systems unlock value far greater than their old-line industry, because they transform the assumptions of their old-school industry. The best hybrid systems forget about the constraints of today, and reconceptualize the consumer value proposition with the hybrid advantages at the core. Here are some good examples:

  1. Uber: At the core, global network planning, dynamic pricing, and GPS-enabled “call a car” buttons transform mobility. Using apps is approximately ∞% better than calling a taxi dispatcher on a payphone, describing your cross-streets, and hoping that someone shows up in the next 10–45 minutes.

  2. Google: it’s easy to forget that everyone who lived until twenty-two years ago did not have access to the sum total of human knowledge. Hopefully, your local library had a book? And the local librarian could help you find it? Google did not replace librarians — it upended the entire idea of finding knowledge, and delivered it to us all (in 100ms or less.)

  3. F-117 Nighthawk: This stealth plane is one of the first to be so aerodynamically unstable that humans cannot fly it: humans can’t react quickly enough to prevent a minor error from literally spiraling out of control. But

  4. Lasik: I can see today because one day, I had precise amounts of my eye burnt off by lasers. That’s impossible without a system that adjusts laser pulses more than 100 times per second.

I call this the Human Law: removing humans from monotonous workflows adds more efficiency than you estimate, even when you take account the Human Law. (I am “adapting” Hofstadter’s Law.) A great hybrid system removes the human cogs, empowers the human leaders, and delivers a service that is a step-change from what came before.


How to not apply AI

I’m an AI alignment realist: Today, AI is a tool that will be aligned with the values of whoever creates it. We create tools that accomplish what we specify, whether those goals are positive, negative, or “use the past to inform the future.”

And that is why a hybrid system is not trying to get rid of humans. Perfectly repeating the past is helpful tactics, but terrible strategy. A great hybrid system pairs the best of computers — an indefatigable executor — with smart, motivated leaders. We need humans who can think about the objectives of a complex system, take responsibility for its effects on the real world (which is inhabited by humans), and adjust the goals. And I believe that once humans are removed from the mental load of monotony, we can make better strategic decisions.

This is part of the investment thesis: where AI can extract greater efficiency, it’s time to use AI. But this investment thesis does not describe the optimal outcome for all systems. In fact, many systems should not be designed for efficiency, because those systems have other, higher goals. For example, focusing on economic efficiency in the criminal justice system would be horrifying. So don’t do that.


Choose Your Own Adventure

  • “I’d like to speak with you about something.” I’d love hearing from you. About me & contact details.

  • “I think this insight should be applied into the field of law!” Yes. Yes indeed. Check out Tome.

  • “I’d like you to invest in me!” I’d love to invest in you. I only invest through SignalFire. Read more about SignalFire.

  • “I think you’re dumb!” Tell me why! In fact, there’s a whole website designed for this: tweet at me.