As a person obsessed with high performance, the idea of creative expression comes up more and more when I think about what great work is really about.

We often think of a person when looking at great work, because great work tends to inherit certain characteristics of its creator. When we look at Tesla, we think of Elon Musk. When we look at the iPhone, we think of Steve Jobs. When we look at Airbnb, we think of Brian Chesky.

Is this another mindless admiration of founders being complimented over and over again? I think not - it is because great work happens to carry the shape of a person’s mind.

Great work is hard to define, but we know it when we see it. If I were to speak to the commonality of great work, I would say this: great work is new knowledge expressed in creative forms. It is not merely execution, productivity, or doing something difficult. Great work reveals non-obvious truth and expresses it in a form that other people can follow along.

There are many such examples, ranging from category-defining products to traditional products done in category-defining ways. Steve Jobs had simplicity baked into his head. He believed the interaction between people and products should be extremely minimalistic. He expressed this knowledge through the delivery of minimalistic stores, something never seen by the world before.

Early Netflix was another strange example. Mailing physical discs to people sounded almost ridiculous, but it contained a non-obvious truth: they value selection, convenience, and control over the in-store experience held dearly by the incumbent then, Blockbuster.

There is often something unusual or weird about great work, especially during its early days, because great work is often very personal. Its creators create it almost as an indulgent personal experience, not caring too much what the rest of the world thinks. Given how indulgent it is, they can often go incredibly deep, digging up unknown truths about what works and what does not, and keep going until they arrive at a self-coherent gold mine.

In other words, great work is the result of someone obsessed navigating through the idea maze tirelessly.

Great work, as the creative form of new knowledge, is even more important in the AI era because AI is commoditising existing knowledge fast. The future may reward people who produce new knowledge fastest, rather than people who hold the largest amount of existing knowledge, as is often the case today.

The important distinction is this: some knowledge can be searched, summarised, and recombined. But some knowledge can only be produced through action. When Elon Musk reflected on the journey of Starlink, he said the team had to invent the technology along the way because the technology needed to support Starlink did not exist at the time. This is indeed incredible.

So how do we do great work?

The only correct recipe for great work is great people.

I am always of the opinion that to find great people, you have to be great first, preferably at something that matters to others, because great people only want to work with other great people. Pretending otherwise is a waste of time.

Being great at something matters a lot because of the bar it sets. The efficient market hypothesis has many forms in the world, including the 80/20 principle. People can spend 20% of their effort to gain 80% of the knowledge, but to get to the remaining 20% of knowledge requires 80% of the effort.

In the business world, we often say this is a waste of time. But I increasingly think this is wrong.

The people who truly get to 100% on something will never be the same again, regardless of the other fields they step into later. They now know what depth feels like, what fake understanding smells like, how laughable shallowness is. I suspect that being relentless in a certain craft is the source of taste.

What’s more important is that once a person is great at something, this person can often spot other great people who may otherwise be invisible or too early. I truly believe in the principle of betting on the hungry over the proven, and this principle has worked far too well for my own team.

One signal of great people is whether they have creative expressions in certain shapes or forms, even when those expressions are not great work yet. It does not even need to be physical or virtual. It could just be a weird and wonderful theory we have never heard of.

I often see Vitalik Buterin, the creator of Ethereum, being mocked for his social awkwardness, appearance, or even broken socks on social media. But his theory around programmable money was weird and wonderful in 2014 for anyone who listened.

Raw signals are often unpolished and low-status. This is probably also why status-seeking people are often shallow.

I do still believe that great people eventually get recognised in this world, in one way or another. In the worst-case scenario where no one else believes in them, great people should still be the biggest sponsor of their own work and find a way out.

However, when we need a team to produce something great, it is often much easier to find great people when they are still early and not obvious, rather than when they are already well known to be great and hard to hire. We get to develop along the way as a bonus point.

Assuming the hard work of finding great people is done, it is still not enough to produce great work in a team setting. The reason is simple: undirected team members cancel each other out.

If this is the concern, should teams then optimise for maximum harmony?

Absolutely not.

Harmony sounds good, but harmony often means full alignment on everything all the time. This forces very different individuals into the highest common denominator of agreeable ideas. The ideas that result from such a setting will necessarily be common ideas with little value, like:

Those ideas aren’t wrong technically, but they are worse than wrong, as they are too obvious.

In fact, what is special about high-performing teams is often their ability to hold contrarian beliefs against popular “truths”.

Of course, most contrarian ideas are simply wrong ideas. Being contrarian alone is not valuable as we also have way too many opinionated idiots. What makes a high-performing team’s contrarian bet right is that they have a shared vision that requires the current reality to change. The current world does not naturally support the vision, otherwise everyone would already be doing it. So the team has to find a non-obvious path through the idea maze, move quickly, and keep the idea alive long enough for reality to start bending around it.

The contrarian part is needed for something big. But the right part requires shared vision, exceptional taste, and crazily fast speed.

An ambitious and contrarian vision happens to correspond to weird and wonderful great people. But generally, it is still dominated by one person’s idea, preached to the entire team to achieve alignment, like it or not. This may sound uncomfortable, but I think it is mostly true. Great work is rarely born from averaging everyone’s opinions. It usually starts with one person seeing something others do not see yet, and creating enough force for others to see it too.

So how do teams become crazily fast? The answer is not collaboration, but ownership.

Everyone owns a non-overlapping piece of the pie that corresponds to their unique strengths. They do not need to agree on everything and do not need to be involved in everything.

In fact, the only thing they need to share is direction and standard. The rest should all be individuals’ creative expression at large.