I find it interesting how the idea of “decoupling” shows that outputs like essays no longer guarantee that real learning happened. It makes me think that focusing only on artifacts has always been a weak way to measure understanding, even before AI. The concept of tracking “friction traces” feels like a more realistic way to capture actual learning.
The line that hit hardest: the artifact was upstream evidence of the process, and generative AI broke the causal chain entirely. The seven friction traces reframe the whole problem — stop trying to detect AI, start measuring whether learning actually occurred. One can't be faked without doing the other
Powerful reflection on the end of the artifact as a stand-alone merit. As generative AI saturates the world with 'perfect' content, we are rediscovering that art was always a bridge between two ghosts in the machine. When the bridge is built by an algorithm, the connection vanishes. We’re entering the age of the Artist, not the Art
I find it interesting how the idea of “decoupling” shows that outputs like essays no longer guarantee that real learning happened. It makes me think that focusing only on artifacts has always been a weak way to measure understanding, even before AI. The concept of tracking “friction traces” feels like a more realistic way to capture actual learning.
The line that hit hardest: the artifact was upstream evidence of the process, and generative AI broke the causal chain entirely. The seven friction traces reframe the whole problem — stop trying to detect AI, start measuring whether learning actually occurred. One can't be faked without doing the other
Powerful reflection on the end of the artifact as a stand-alone merit. As generative AI saturates the world with 'perfect' content, we are rediscovering that art was always a bridge between two ghosts in the machine. When the bridge is built by an algorithm, the connection vanishes. We’re entering the age of the Artist, not the Art