The Gatekeeper’s Dilemma

I am a product manager. Lately, I was reflecting on the nature of my role. The role did not emerge out of thin air. It was invented to manage the friction of the previous technological generation. Product managers sit between the people who want a thing built and the people who understand the convoluted machinery required to build it. We exist to navigate complexity.

But I am starting to realize that the very complexity my profession was built to manage is evaporating in front of me.

The internet democratized distribution. It gutted newspapers, transformed media, and forced knowledge work to chase global labor arbitrage. Yet through all that disruption, technical gatekeepers held their ground. They survived because they controlled the complexity moats. If you understood the arcane syntax, the undocumented legacy systems, or the gnarly deployment pipelines, you held power.

Now, large language models are coming for those moats. They act as universal complexity busters. They possess the ability to destroy the barriers that protect the initiated. This will cause chaos, and my own role is likely no longer relevant for the work it was originally designed to do. But chaos is also opportunity. The impetus to build remains. As long as that reward exists, new monuments will rise to surpass the old towers. We are seeing this in real time. New companies are forming at insane private valuations. While some of it is a bubble, real value accretion is happening if you look closely. I have to evaluate this without a moral lens. Defending the status quo requires accepting the strange premise that the current arrangement is inherently right. It is not. We have to take morality out of the equation to see what comes next.

The Moat of Formats

Complexity is rarely an accident. I was recently reading Mat Duggan’s analysis of the origins of the .docx file format. Microsoft did not make the format gnarly because they lacked engineering talent. They made it complex because that complexity served as a moat. Controlling the file format protected the business and locked users into the ecosystem. The enterprise sales relationship became the ultimate defense, and the technical complexity was the wall around the castle.

But imagine trying to build that kind of moat today. In a world shaped by the internet, open source software, Google, and now LLMs, artificial complexity cannot hold.

I experienced this collapse firsthand. For years, I paid for WordPress. I loved the platform, but I wanted the clean simplicity of a static site generator. The barrier was not just technical debt. It was the complexity of the migration, the fear of breaking my existing site, and the reality that I had limited time. I knew the project would be spread too thin across weeks. I wanted to understand the process so I could learn, but that friction kept me paralyzed.

Then, I sat down with an LLM. It did not just dump code on my screen. It acted as a translator. I prompted it to explain the migration in a way that built my confidence. It broke down the process, reduced the complexity, and compressed the timeline. The moat evaporated, and I migrated the entire site in a single day. The model generated the unlock.

We are watching companies try to manage this shift in real time. Anthropic is aggressively protecting its agent harness in Claude Code. Meanwhile, new platforms like Pi are emerging to break you free from those very harnesses, allowing you to seamlessly route work across Claude, Codex, and Gemini. The lesson is clear: technical complexity moats are no longer durable. At the current pace of development, even possessing a “frontier” model is not a lasting moat.

The Evaporation of Gatekeeping

If technical complexity is no longer a durable moat, we have to ask what happens to the industries built entirely on artificial friction. In an infinite timescale, artificial complexity approaches zero.

Look at the legal profession. The primary moat is obscure legalese and convoluted regulatory codes. It is fundamentally a language barrier. When models can translate, verify, and generate legally sound documents perfectly, the traditional gatekeepers lose their tollbooth. This means that lawyers, people currently paid to manage the complexity, are going to need to define their value or change their rates to meet the market. My prediction is that there are going to be a massive number of tools that will help plaintiffs represent themselves in new ways as the destruction of complexity meets the creativity that is currently gatekept.

Look at healthcare and insurance. The administrative system relies on an adversarial moat of billing codes and opaque claims processes. Patients remain locked out because the system is intentionally illegible. But LLMs are uniquely suited to decode that exact type of structural density. They can map the codes, route the appeals, and dispute the denials at scale. This is one that I am very excited about. I am currently paying $1,500 a month in insurance premiums just to get the permission to spend another $20,000 deductible. That is insane. What is the value that I am getting for this? I receive some potential discounts for very limited medical procedures and almost no backstop for a medical emergency given that high deductible. I cannot wait for this to be disrupted.

Look at the tax code software. Entire sectors exist solely to guide clients through artificial mazes. What happens when an intelligence layer can navigate accounting loopholes or an SAP migration better than an implementation consultant?

In quantitative finance, the knowledge, access, and application automation is ripe to accelerate the democratization that has already started. When regular people gain a tool that can instantly translate the complex into the accessible, the power dynamics invert.

The destruction of these moats is going to displace a massive amount of white-collar gatekeeping. We assume these verticals are inherently complex, but they are often just artificially guarded. We simply lacked the right translator. Now, the translator is effectively free. And the walls are coming down.

This all makes me bet on the open web more. When complexity becomes democratized, there is very little value in protecting and gatekeeping knowledge. Maybe I am biased and I am viewing this only through a lens of what’s still available in fron of me. It becomes futile or at least time-constrained. That means people will be motivated to share, either to gain credibility or to monetize the process of sharing. This makes me bullish on open source and the open web. Who knows, maybe this will convince others to also put their name to their thoughts and build their own audience.