Your objective versus subjective line explains something I keep seeing inside companies as well. Nobody objects when AI reconciles the data or drafts the meeting summary. The resistance starts when AI writes the message that carries someone's judgment, because the reader is no longer pricing the information, they are pricing the person. That is why the button is really a norm-setting tool rather than a detection tool. Detection asks whether a machine wrote it. The norm asks whether anyone stands behind it, and that is the question that has been missing.
I think it’s less about whether it’s actually AI-generated or not and more “this is a bunch of LinkedIn BS.” It’s a way of sending a signal that a piece of content is low quality for the algorithm to take into account, plus LinkedIn will share with people that others think they’re posting inauthentic crap. Ultimately, I think that delivering feedback on what strikes people as inauthentic garbage will be a good thing. More than anything, I’m happy that LinkedIn is trying SOMETHING,
Two approaches to AI slop and both bad. You can't do anything now about people marking you as AI slop on LinkedIn, I wrote about this https://thefunnelisdead.substack.com/p/linkedins-seems-like-ai-slop-feature back on July 31. But you can do something about Pangram's mistakes on your content in Substack. Just write to Substack's privacy department and ask them to give you account-level opt-out for Pangram. It helps if you are in the EU, or in California or a handful of other states that have privacy laws on the books. Pangram has apparently been turned off for EU customers due to all the backlash.
Your objective versus subjective line explains something I keep seeing inside companies as well. Nobody objects when AI reconciles the data or drafts the meeting summary. The resistance starts when AI writes the message that carries someone's judgment, because the reader is no longer pricing the information, they are pricing the person. That is why the button is really a norm-setting tool rather than a detection tool. Detection asks whether a machine wrote it. The norm asks whether anyone stands behind it, and that is the question that has been missing.
Exactly! I think that's the line we'll see develop on what we want to come from AI vs what we don't.
Chris Best’s dig at LinkedIn is still my favorite part about all of this.
Ha me too
So good and I don't know how anyone can argue with him after that line.
Interesting post on objective vs subjective AI output.
Will no one speak for Medium?
Hahah. Medium is now pretty much nothing but engineers writing heavily AI-generated posts. It's a weird, special place and I think they're good?
I feel like Medium earned the distinction of being the first platform to be overrun by AI slop, and they don't get enough credit for that innovation.
True trailblazers
People are terrible judges of whether content is AI-generated or not. What could go wrong?
I think it’s less about whether it’s actually AI-generated or not and more “this is a bunch of LinkedIn BS.” It’s a way of sending a signal that a piece of content is low quality for the algorithm to take into account, plus LinkedIn will share with people that others think they’re posting inauthentic crap. Ultimately, I think that delivering feedback on what strikes people as inauthentic garbage will be a good thing. More than anything, I’m happy that LinkedIn is trying SOMETHING,
Agree that some action is better than none.
Two approaches to AI slop and both bad. You can't do anything now about people marking you as AI slop on LinkedIn, I wrote about this https://thefunnelisdead.substack.com/p/linkedins-seems-like-ai-slop-feature back on July 31. But you can do something about Pangram's mistakes on your content in Substack. Just write to Substack's privacy department and ask them to give you account-level opt-out for Pangram. It helps if you are in the EU, or in California or a handful of other states that have privacy laws on the books. Pangram has apparently been turned off for EU customers due to all the backlash.