Taste as the Last Moat
SI has collapsed the cost of execution. That changes what the game is. An argument for why taste, the judgment that knows what is worth making, becomes more valuable when everything is easy to make.
SI, as used in this piece, stands for Super Intelligence.
Somewhere in the last two years, the cost of making things fell through the floor. A video that once required a production crew can now be roughed out from a prompt in an afternoon. Websites, documents, presentations, code: the cost of producing a first version of any of them has dropped sharply.
This is good news. But it creates a problem that most people in the technology industry are reluctant to state plainly: if everything is easy to make, the question of what to make becomes the whole game.
Taste is the capacity to answer that question well. And despite what you might expect from a company building SI products, it is becoming more valuable, not less.
A clarification
Taste is not the same as style, though the two are often confused. Style is preference: this font rather than that one, this palette over another. Style is relatively easy to copy and easy to chase. A stylish company can lose its style the moment it hires the wrong creative director or starts shipping features by committee.
Taste is something harder to name. It is the judgment that knows which ideas to kill before they reach the brief, which version of a thing is finished and which is still wrong. It operates before the work begins and after the work is done. You can have terrible style and excellent taste. You can have impeccable style and no taste at all.
The person with taste knows when to stop. This sounds simple. It is not.
What SI changes
Before SI, the bottleneck in creative and professional work was execution. You needed to know how to use Photoshop, or how to write code, or how to format a legal document. Those skills took years to acquire and created real barriers. Not in every field, but in many. The tasteful amateur was often blocked from expressing a vision by the gap between their standards and their technical capacity.
SI closes that gap substantially. Not entirely. Execution still matters, and it matters more in some domains than others. But it closes it enough that what to make becomes a far more binding constraint than how.
And here is what SI does not change: it does not know what is worth making. It can generate a hundred variations of a pitch deck or a film treatment, but it cannot tell you which one is right for this client, this moment, this story. It does not feel the difference between an idea that merely works and one that lands. That judgment requires a person. Specifically, a person who has spent years developing the calibration to feel that difference reliably.
Why this matters here
Every venture in this group works in a domain where taste is the input you cannot skip. Film is the clearest case: something that looks technically correct but feels emotionally hollow is not a film. It is a rendering. The same principle runs through art and spatial design, and through travel and real estate analysis. In all of them, the gap between getting the technique right and getting the feeling right is exactly what SI exposes rather than closes.
We are not building products where the model or the data pipeline is what sets them apart. Those are commodities, or close to it. We are building in places where the question is whether the output feels right: to a collector examining a certificate, to a director reviewing a cut, to a traveler opening an itinerary. That standard is not set by a benchmark. It is set by someone who has learned to feel the difference.
And that person is very difficult to scale. You can hire taste, for a time. You cannot manufacture the decade of forming opinions in good rooms, paying close attention, being wrong and slowly getting it right.
The irony is that SI makes this scarcity more visible. When everyone has access to the same generation tools, what you produce reveals exactly what you were trying to make. If you did not have a clear idea of what good looked like, the output will show it.
The actual argument
This is not a case against SI. It is a case that the people who will build the most interesting things with SI showed up with taste before the tools got cheap.
That is what Gilded Edge Group is. A collection of ventures built by someone who spent years forming opinions about what quality looks like in film, in art, in travel, in design and in documentation, before there was a model that could help carry those opinions out. The tools speed up the making. They do not set the standard.
The moat is not the technology. The moat is the answer to the question, posed clearly and early, of what is worth making. That question does not get easier to answer because the making gets cheaper. If anything, it gets harder.