So, if instead of saying “I am observing a conversation between two pundits about the most effective ways to influence the broader American public” you say “I am offended that this person insulted me by saying I need to make videos or else I hate democracy” you aren’t as text based as you thought.

2) audience - Matty and Jamelle aren’t addressing people like me, who do not have or hope to have any influence. I’ll never make videos but that’s fine because nobody cares what I think. They’re only talking to and about journalists or influencers or the substack set.

1) context - I believe the context of Jamelle’s post is a Matty Yglesias post about how Bluesky users should go back to Twitter to engage with their users and try to persuade them, with the implication being that this is how we will persuade the electorate more generally.

“AI is a tool of fascism designed to destroy organized labor, with environmental destruction and child suicide as a byproduct, but at least it treats me like I’m smart for asking a question instead of stupid for not already knowing.” We AI haters should not underestimate the power of being nice.

I also think it is somehow connected to the complaint I sometimes see from political moderates that “the left is so much MEANER than the right” interpersonally - that sure the republicans are evil but at least they don’t nitpick your politics to death.

The hackey sack guys and the ultimate frisbee guys met again and made peace with one another on the disc golf course.

Well, I guess I didn’t hold back the urge I’m SORRY I REALLY TRIED!!!!!!!

Ok here’s the real problem with AI

When you start parsing what work is unimportant enough for the AI to do and what only you can do, you are making the assumption that you know what parts are “important.” This is unknowable opportunity cost. What discoveries would you have made writing the outline?

I still have the sign, actually.