100% with you on this. I wrote more than a year ago about A8 detectors, how they are doomed to fail for the same technical reasons that LLMs work, and why they are asking the wrong question to begin with, as you said. Plus, the hypocrisy of Substack in turning to AI detectors as a way to attest our humanity. Don't even get me started on that... Great post, man.
"....You don’t build a detector for people you trust.
Substack’s value proposition was to build a social space based on human relationship, not algorithms — a network built on people choosing to read each other, not a machine-curated feed.
A detector doesn’t fit into that idea. It quietly replaces it with suspicion as the default posture, whether or not anyone ever gets flagged......"
While some of this is true, also true is that when money is a motivator, well then cheating on the inputs and mailing it in so to speak, quickly comes into play. This has always be the challenge with journalism, or what people these days call "content". Money corrupts it, whether it is from advertising or from subscriptions.
Agreed. The problem really isn't AI ... its the incentive structure we've built.
Honestly, I never really saw a problem with AI slop on Substack, because it is built around relationship. So its solving for a problem that doesn't really exist in my mind.
If you don't like what someone is writing, AI or otherwise, don't subscribe!
100% with you on this. I wrote more than a year ago about A8 detectors, how they are doomed to fail for the same technical reasons that LLMs work, and why they are asking the wrong question to begin with, as you said. Plus, the hypocrisy of Substack in turning to AI detectors as a way to attest our humanity. Don't even get me started on that... Great post, man.
"....You don’t build a detector for people you trust.
Substack’s value proposition was to build a social space based on human relationship, not algorithms — a network built on people choosing to read each other, not a machine-curated feed.
A detector doesn’t fit into that idea. It quietly replaces it with suspicion as the default posture, whether or not anyone ever gets flagged......"
While some of this is true, also true is that when money is a motivator, well then cheating on the inputs and mailing it in so to speak, quickly comes into play. This has always be the challenge with journalism, or what people these days call "content". Money corrupts it, whether it is from advertising or from subscriptions.
Agreed. The problem really isn't AI ... its the incentive structure we've built.
Honestly, I never really saw a problem with AI slop on Substack, because it is built around relationship. So its solving for a problem that doesn't really exist in my mind.
If you don't like what someone is writing, AI or otherwise, don't subscribe!