Claude does pick the topics! I worry I did too good a job steering it. I found a few of them insightful too, but it doesn't matter, what matters is that they're same old "They were doing a thing. It failed." Claude writing you see everywhere, intensifying the feeling that nobody on LinkedIn has a pulse.

On the bright side, if everyone is burning credits mining generic but helpful knowledge out of the LLM training data in bite-size packages, I don't need a subscription of my own!

It will be really fascinating to see what happens with the potential for LLM-generated output to become a dominant component of new training data.