A microscopic video wins a competition, then gets disqualified because the photographer used AI to enhance it—and suddenly the line between documentation and fabrication isn't where we thought it was. Meanwhile, a researcher building a writing tool deliberately constrains what AI can do, because unlimited assistance makes thinking invisible to itself. The mechanism is the same in both: we've created systems that can produce perfect-looking outputs, but we've destroyed the feedback loop that lets us know if we're seeing reality or a simulation of it.
When a tool can make anything look professional, judges and users stop trusting their own judgment and start requiring external permission structures—competitions need rules, writers need constraints, institutions need audits—because the surface no longer signals the truth beneath it.
2
Dr. Ning Xu's disqualified video and Jim Lau's deliberately limited AI both point to the same trap: the easier it becomes to make something, the harder it becomes to know whether you actually made it, and that gap between making and knowing is where institutions gain power they never had before.
3
The next winners and builders will be the ones who voluntarily make their work harder to produce—who leave visible traces of their thinking, who constrain themselves not because the rules demand it but because constraint is now the only reliable signal of authenticity.
We don't fear the tools that replace us; we fear the tools that make replacement invisible, and so we invent gatekeepers in desperation, which is simply another way of admitting we no longer trust ourselves to see.