The Realization
I came up in web3. I have spent real years inside the dream of permissionless, trustless, decentralized systems that route around middlemen and let humans coordinate at internet scale without a king.
I still believe in that dream as a long-arc target. And I have come to believe, candidly, that for almost any real disruption inside the next decade, centralization is the right answer.
That is a heel turn from where I started. It is also, I think, true.
The Hard Truth About Trustless Systems
Designing a permissionless, trustless system that is also highly functional, fast, evolving, and beautiful to use is orders of magnitude harder than designing the equivalent centralized version.
It is not impossible. We have proof points (Bitcoin, Ethereum, a handful of others). But the cost in coordination time, in protocol overhead, in the loss of design velocity, is enormous. And the quality bar industries actually demand is not "eventually consistent in 12 minutes." It is "ships a new version this week, fixes the edge case by Friday, anticipates the next one before it lands."
A trustless protocol cannot move at that speed. Its dependencies are governance processes, validator sets, social consensus. Those are the wrong substrates for keeping pace with how fast society is now evolving in the AI era.
For a serious operator trying to deliver win-win value into a fast-moving industry, the right architecture is rarely a DAO. It is usually a small, high-trust team that can ship.
How Platforms Actually Win
People end up on platforms for one of two reasons:
- The utility is so high they would put up with anything to use it.
- They trust the person (or team) running it.
Almost every successful platform in history is one or both. Decentralization is not on the list. It is occasionally a marketing surface, but it is almost never the load-bearing reason a real human commits to a platform with their data, their money, or their attention.
The platforms that genuinely disrupt entrenched industries are the ones that can deliver so much functionality that the alternatives feel obsolete. That level of functionality requires centralization, because every dependency at that level needs to be evolving as fast as the team is. Trustless dependencies almost never are.
Why I Am Now Bullish On Centralization
Centralization gets a bad name because most of the time it is run by people who optimize for extraction. Big-tech surveillance platforms. Walled-garden app stores. Banks. The pattern is so consistent that "centralized" became a synonym for "rent-seeking and predatory."
That association is doing a lot of damage. It is making serious builders refuse to centralize when centralization is exactly what their would-be users actually need.
The frame I have landed on: centralization itself is morally neutral. What matters is who is doing the centralizing and what they are pointing the centralized power at.
If the centralization is in the hands of a person whose actual orientation is service, who has a track record of building win-win systems, who refuses extractive defaults even when they would be profitable, and who treats every user as someone they are accountable to: the centralized version of that platform is better for everyone than any decentralized alternative within reach.
This connects to heartshare and to inner circles. Trust is the asset. The architecture follows the trust.
The Chosen-One Condition
Here is where this gets faith-shaped, because it is.
I think the world right now is desperate for chosen ones: human unicorns who can assemble chosen teams to evolve and elevate entire industries, with a clarity and a covenant that pure economic rationality cannot generate on its own.
Not "chosen" in a self-congratulatory sense. Chosen as in called. Chosen as in operating under a real mandate, with a real conscience, with a real accountability to something larger than the cap table. Chosen as in a person whose centralization of an industry's most efficient new value-delivery mechanism is better than the alternative because they will not abuse it.
That bar is so high that almost no centralization currently meets it. Which is why so much existing centralization is rightly distrusted. The answer is not to abandon centralization. The answer is to raise the bar on who gets to centralize.
The centralization of a brand new, most-efficient way of delivering value in a win-win capacity for players within an industry should be done by the chosen people, with a divine mandate. Said directly because it is the truest version of what I think.
When Decentralization Is Still Right
This take is not anti-decentralization. It is anti-default decentralization.
Decentralization is genuinely correct when:
- The asset must outlive any single team or country. Bitcoin's whole point.
- Capture by any single actor would be catastrophic. Public infrastructure, identity, money in jurisdictions where the central authority is hostile.
- The value is in the openness itself. Open standards, open protocols, open knowledge.
For those, decentralization is the right architecture and centralization would be the abuse. I still want a sovereign substrate for everything that matters at civilizational scale. See progressive sovereignty for the longer arc.
But for the next ten years of disruption inside specific industries, the answer is almost always: a chosen team, centralizing fast, refusing to extract. The decentralization conversation can resume after the legacy systems they are replacing have been broken. Right now the priority is breaking them.
What This Means For Builders
If you are building something disruptive and you have been hesitating to centralize because the web3 voice in your head told you not to: consider that voice may be wrong for what you are building.
Ask the harder questions instead:
- Are you the right person to hold the keys? (Honest answer.)
- Is your team the right team to evolve with the user base? (Track record.)
- Will you refuse to extract even when extraction would be profitable? (Test it on a small thing first.)
- Are you accountable to something larger than your investors? (You should know.)
If all four are yes, centralize without apology. If any of them is no, do not centralize. Either upgrade your team, upgrade your mandate, or step aside for someone who can clear the bar.
Centralization in the hands of the chosen one, on a divine mandate, is the fastest path to disruption that actually serves the people being disrupted into. Decentralization is the right answer for what must outlive us. Both can be true. Default to the right one for the work.