Jumping the Valley: How to Innovate Like No Other

Part 1

Pick a game you can win.

Again, pick a game you can win.

You may think work = success, like a linear equation. The more work, or investment of any sort, the more success.

But that isn’t reality. Here’s an example: the richest person in the world (Jeff Bezos) is worth $160 billion. I bet Jeff Bezos works pretty hard. But it is simply impossible to comprehend $160 billion. Some back-of-the-envelope math: the average American would have to work 3 million years to accumulate Bezos’s wealth. It seems impossible. But that is the type of success I will be talking about here. And once you understand the mechanics behind success, it will start to make sense.

Rather than being linear, the relationship between investment in an enterprise and success is often exponential, like this:

Screen Shot 2018-10-11 at 12.34.08 AM

Not every enterprise will have a graph that looks like this. But you want to seek out one that does because the gains are so great (if you play your cards right).

It makes sense to seek out enterprises with exponential, compounding returns to investment, and specialize in those.

Imagine three industries where success = work squared. You have 10 units of work to spend. Your work, spread across 5 such industries yields 20 units of success. Your work, specialized in one industry, yields 100 units of success.

Specialization is an important principle. But the one principle, more important than anything, is:


The significance of monopoly is almost impossible to understate. All of my work thus far has been building to this point.

Peter Thiel once wrote, “There are two kinds of businesses. Monopolies, and those that never make any money”. In one of the most fascinating interviews I’ve ever watched, Thiel’s interviewer, aghast, asked him if he was being facetious, or perhaps saying it for effect.

“No”, answered Thiel.

These videos changed how I viewed economics and business forever:

Another quote (paraphrasing)

There is this somewhat strange phenomenon in Silicon Valley where so many of the successful companies seem to be created by people who are suffering form a mild form of Aspergers. We need to turn this around as a critique of our broader society: what is it about a society where anyone who does not have Aspergers is at a disadvantage because they are talked out of their original, interesting ideas before they’re even fully formed? “That’s a little bit too weird”, they say to your ideas, people look at you funny, and so you think you should do something more conventional.

If you want to see just who extreme competitive forces are, look at video game speed runners. Look at the such extreme lengths they go to push the boundaries of human capabilities, to compete to be the best. Seeing the lengths that people go if video games, just for fun, imagine the lengths that people go to in industries where money is one the line. If you are in a perfectly competitive industry, you are competing with these types of people to differentiate, the absolute best in the world. Unless you have godly, innate ability, and decades to get better, you will always loose.

If you have taken Microeconomics, you have seen these graphs:

monopoly and competition
In perfect competition, where there is supernormal profit to be made, firms will enter the market, until price drops to a level where supernormal profits are erased. In monopoly conditions, by contrast, marginal revenue and average revenue are downwards sloping because the single firm controls the quantity supplied (and by extension, the price). The firm can simply choose whatever price and quantity supplied most benefits them, given their costs. And there is nothing consumers can do, because they have no alternative choices. The PRSC rectangle shown in the graph represents what I call the “rent” (supernormal profit).

If you have no monopoly, you are a commodity. You are compensated for your mere time, (your most valuable commodity), and nothing more. You have no supernormal profit, no excess/passive income, you collect no colloquial rent.

But I don’t want to found a business! I don’t want to be rich!

It doesn’t matter, the principles I am outlining here apply to success in all enterprises.

Microeconomics 101 also teaches about oligopoly and monopolistic competition. For all intents and purposes, those don’t exist in the real world. Thiel emphasizes that there is shockingly little in between monopoly and perfect competition, which that is the key divide in business. The difference is blurred because both sides practice obscurantist rhetoric: monopolies try to define their industry broadly, so as to avoid anti trust, whereas perfectly competitive business try to define their industry narrowly, so as to sound attractive to customers and investors. How you define the scope of the industry is a foundational semantic question.

Which brings us back to the title of the post. It is far better to be the “best” at a small but rarely played game, as opposed to second place at a popular game.

In this context, game refers to any sort of competition (in contrast to my other post, where I analyze the literal meaning of the word).

Pick a game obscure enough that you can win. And the the easiest game you can win is one where you are the only player. In business, you should create or reinvent a new industry, where an industry is a “game”. Mark Zuckerberg reinvented the social network “game”. Steve Jobs reinvented the mobile phone “game”. For a time, they were the only “players”, or industry leaders.

Monopolize small markets first, before going after large markets. Small markets are easier to dominate.

Which leads us to the next question: How do you create a new game?

I can give you a few guiding principles.

1: Understand the Types of Monopolies

Monopolies usually fall into one or several of just a small number of categories. This video is a good summation.


Build monopoly out of brand loyalty through consistency and standards.

Proprietary Technology

In order for proprietary technology to grant monopoly, it must present a difference in kind rather than a difference in scale from the existing technology (scale means speed, cost, etc.) With one exception: if a difference in scale is 10x better than existing technology, it rolls over to being a difference in kind.

The link takes you to an Extra Credits (game design) video, but the principle is more general.

Copyright Protection

Copyright protection and proprietary technology are twins; of course it’s necessary to patent technology to keep a competitive advantage. “Copyright” though has a more expansive connotation. For example, J.K. Rowling and her publishers have a monopoly over the Harry Potter books, so she made a fortune.

Network Externalities/Network Effects

Explained here.

Zero-sum resources

Explained here.

What that post does not fully explore is right-of-way (transportation infrastructure) services like utilities, railroads, and so on, which benefit from affording people and resources to access across land. I will delve into those in a separate post.

2. Plan, don’t Roam

A common theme on this website is the distinction between what is designed vs what is evolved. “Evolved” can refer to biological evolution, but I also use it to refer to a certain trial-and-error process.

Consider how a school of fish decides where to swim. Each fish is allowed to swim in its own direction, as long as it doesn’t swim too far away from the school. The choices of hundreds of individual fish culminate into a collective decision about where the school should swim. By this process, a school of fish moves slower and less decisively than an individual fish can, but the swim direction is a sort of “collective decision”. It’s the same type of situation with a herd of roaming buffalo.

Now consider how a large organization operates. An organization has hundreds, thousands, or even tens of thousands of individual actors. They are each working to some degree independently, and to some degree in parallel with the organization. There is a little bit of freedom for people to experiment on the edges. If an experiment works, there is a domino effect of others following suit, until the entire organization aligns with that objective. But it is a slow process.

This is the main reason why small organizations have the reputation of being more efficient, and more innovate, while larger ones are more bureaucratic and stuck in their ways.

A large organization, like a herd of buffalo, engage in what I’ll call “roaming”, a process of gradual evolution through experimentation of those individuals on the edges. You don’t want to do that, because blind-luck experimentation only allows change in small jumps.

You want to plan, or build from the ground up with a unifying vision.

Very few actions are actually planned, how I define it. Many actions give off the appearance of being planned, because at some stage a decision maker signed off on them. But, more often than not, the decision maker didn’t have a “choice” in the classic sense; they signed off on the obviously right solution because it was the way the world was going (whether they like it or not). Either that, or the choice could be attributed to causes outside of their control.

To understand why you have to plan, I have to reference the adaptive valley:

Screen Shot 2018-09-22 at 5.22.20 AM

The X-axis represents variability in a trait, strategy, or system. The Y-axis represents the utility, or benefit of that strategy.

Now suppose, as is often the case, you find yourself in the following situation. People start out on the very left of the graph, and over time, through endless trial-end-error, they improve efficiency until they get to the top of Peak 1 (where the arrow points).

Screen Shot 2018-10-11 at 11.48.38 PM

Above depicts the current situation. You are where everyone else is. You don’t want to stay there, because the situation is too competitive, too crowded. All the profits are being competed away.

In order to attain monopoly, you will have to (A) get a result that is better than anyone else, and (B) occupy space that no one else is occupying.

What has both of those properties? Peak 2. No one else is occupying it because it requires them to cross “the adaptive valley” (valley 1) – getting worse before they get better, which roamers are not equipped to do.

You will most likely miss peak 2 unless there is a coordinated effort to aim for it. In theory, it’s not impossible to get to peak 2 by accident. However, it’s harder than the oversimplified above graph gives off. In real life, the “horizontal dimension” is actually a complex combination of factors that are to some degree irreducibly complex.

Of course there are other successful planners. We call them innovators. Rest assured, there will always be more peaks in the landscape to explore.

3. Rapid Progress, not Incrementalism

Trial-and-error goes hand-in-hand with incrementalism, because trial-and-error almost always builds of what existed before. It’s a natural-selection-like process. You try a bunch of different small adjustments, keep what works (ditch what doesn’t), and repeat the process continuously.

Rapid progress is quite different. It is achieved through deliberate planning and leadership. Thiel advises rapid progress over incrementalism.

This is an incremental change, shown on the adaptive valley graph:

Screen Shot 2018-10-12 at 12.26.46 AM.png

Any small change will necessarily make you go down. If a small change could work, someone else would’ve figured it out already.

This is a large change:

Screen Shot 2018-10-12 at 12.27.49 AM

You have broached uncharted treasure.

Trial-and-error incrementalism is generally very good at optimization problems, but plenty of people have done that already. You want to do something proprietary.

Consider political discussions. We see too many questions like this:

  • Should we increase social security spending growth by 2% or 3% next year?
  • How should we tinker with the various tax incentives?
  • What non-traditional forms of marriage should the government recognize?
  • Should we add a regulation on the maximum number of patients a nurse can care for in federally funded hospitals?
  • Should immigration be increased or decreased by 3%?

And not nearly enough questions like this:

  • Should we eliminate social security and replace it with a requirement that people put a percentage of their income in annuity?
  • Should we eliminate all tax write-offs/exemptions/credits/loop-holes/deductions, and start from scratch?

Granted, my examples show a libertarian bias. However, I do think that the discussion around UBI, which I have a post about, is an attempt at the latter type of question.

4: First Principles

Elon Musk explains how to buck a trend: extrapolate from first principles.

Most people extrapolate instead from work that has been done before. You can get a leg up on these people going back to basics and asking “why, why why do we have to do things this way?” until you reach fundamentals.

Thiel gets at first principles by asking “crazy questions”, questions so radical they make you re-think the entire dynamic of your strategy:

“How can you achieve your 10 year plan in the next 6 months?”

“How could you 10x your output?”

And so on.

You may be inclined to say, “that is impossible”. The fact that you think that is the point. Thiel does not accept that answer. Challenging the supposedly impossible is how you innovate.

There is a time and place for honoring tradition. But this post is about how to do something new. What I am suggesting can land people in 2nd tier thinking, but only if they copy trendy, popular contrarianism. If they stick their personal, private contrarianism, first principles can get people to 3rd tier thinking (see the link).

5. The Skill Stack

Cartoonist Scott Adams argues what superficially appears to be the exact opposite of what I am arguing, but I will explain why our views do not contradict, and are in fact mutually compatible.

Adams names a skill “stack”. He gives himself as the example. He’s not the funniest guy, or the best drawer, or the best office worker, or the most dependable person. But he is above-average at all of those skills. The combination of unique talents makes him successful.

Adams gives Donald Trump as another example. Trump isn’t the best at insulting, or the most persuasive person, or the best at branding, or the best talent delegator, but he is above-average at all of those, the combination of which made him a very successful campaigner.

When I took the IQ test as a kid, my parents were confused at the result. My “overall score” was higher than my score on any individual subtest. But my score was correct. How could this be?

It is possible because the overall score isn’t an average of the subtests, it’s based on the probability of the combination achieved. Some people perform well on a specific subtest, but not so well on the others. The existence of such people brought my score down on the subtests. It turns out getting a good score on a just one subtest is common, and not that impressive. The fact that I did pretty well on every subtest was impressive – enough to raise my overall result

(They didn’t tell my my IQ, so I can’t brag about it. But I will brag about the fact that I remember that random moment with the psychologist from over a decade ago.)

To succeed, you want to develop a skill stack. But wait… doesn’t that contradict what I said earlier about specializing?

Well, no. The essence of the confusion is the distinction between a skill and a game.

“Game” in this context refers to any sort of competition.

“Skill” refers to the abilities that contribute to success in a competition.

For example, football is a “game”, requiring the skills of running, throwing, catching, (etc.)

Taking tests is a game that requires memorization, improvisation, critical thinking, diligent study, conceptual understanding, (etc.)

Animation is a game that requires drawing, understanding of, and sometimes: story telling, sound design, world building, (etc.)

An election is a game. A business venture is a game. The components which contribute to success at those games are skills.

The distinction is easy to miss because, in common speech, we use the term “skill” to refer to what I’m calling “games”. We say that playing football, taking tests, and animating are “skills”. This is the time to bring up that any skill can be sub-divided into smaller skills.

Some games are virtually synonymous with certain skills. For example:

  • The 100 meter dash
  • Rowing
  • Javelin throwing

These games very hard to dominate for precisely the reason that you cannot leverage a unique skill stack; you must be singularly good at that task. But even in such cases, there are sub-skills, like the steps to throwing a javelin.


Now here is the key point that ties Scott’s premise back to my premise:

If you have a unique enough skill stack, and you create a game specific to that skill stack, you can dominate that game.

Pick a game obscure enough that you can win. And the the easiest game you can win is one where you are the only player. In business, you should create or reinvent a new industry, where an industry is a “game”.

See my point about IQ. It is quite rare to be single best person at a narrow skill. It is much more common to be good at a particular combination of skills. To get to monopoly, you should aim to be the best, which is must more achievable with a skill stack.

So yes, specialize in games. But have many skills.

My advice may seem like platitudes,

“Leverage your strengths!”

“Don’t believe anything is impossible!”

“Don’t copy everyone else. Be radically different and innovate!”

“Keep a unifying vision!”

What can I say, these ideas are memes for a reason.

That’s how you achieve monopoly: pick a game you can win.

Part 3

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s