I was recently invited to join a “prestigious expert panel.”
I had not heard of it.
Since I am oblivious to the world in general, my having not heard of it doesn’t mean much.
Structurally, it looked much like LinkedIn’s pulse/advice section.
Eventually, it became clear to me that they were selling lead gen rather than whatever I had naively thought they were offering, so not for me.
But it raised some interesting questions:
- Given the number of questions per year they host and the number of clients (whoops — I mean “experts” they had), how many answers per year could a particular client expect?
- How proud or anti-proud (embarrassed) would I be to have my answers there?
I didn’t think they would easily tell me how many clients they had, but we could measure the frequency distribution of the experts. I didn’t get that far.
I asked a number of people whose opinions I respect to look at a sampling of their answer postings and give me their assessment of the quality.
Unfortunately, almost everyone I asked gave up because either their ad blocker blocked the content or the ads prevented them from actually reading the examples.
I couldn’t easily read them either.
On forum sites like Reddit, there is a long history of complaints about the readability of that site. (And various other complaints about the organization’s practices.)
I had imported the sample into notion and was, in that manner, able to read them.
They were all true, but none insightful.
I gave ChatGPT-4o the same questions and asked for the same number of suggested answers.
In each case, the Chatbot answers were better.
In evaluating content for publication, I had been saying the question isn’t whether a Chatbot wrote something but rather whether the writing was like that of a Chatbot.
I am now raising the bar:
In evaluating content for publication, is it more worth reading than what a Chatbot would have produced?
I think I’ll decline the honor of being part of that panel.
