I’ve been editing my last post in my head every day since I wrote it.

The ending is too abrupt, I decided. Pieces like this from good writers usually have a third act about some sociological theory that contextualizes for the story. I tapped down notes:

  • Haidt says liberals prioritize care/fairness, conservatives value care/fairness/loyalty/authority/sanctity equally
  • Low-trust world substack very good and relevant not just for thinking about America but also what I should prefer at work
  • Liberals being more neurotic (this is important right)

Then on Wednesday afternoon, a… Coup happened?

At this point in the draft, I wrote down this next line word for word. Then I searched Twitter and of course, someone else had gotten to it first:

It turns out the water I swim is so thick I can’t even form original thoughts. On Christmas Day, I ate a bag of mushrooms in the middle of the desert near Joshua Tree, and I haven’t been able to stop thinking about how we’re nothing more than the sum of our inputs. The world spirit moved through both Alex Halpern and Ian Chan to form the words “sparkling insurrection”, except it moved through Alex Halpern nine hours earlier.

On Twitter, I watched as the world spirit formed a consensus. The people who had stormed the capitol had been behaving rationally given their belief set: the election had been stolen (which some of the people I’d spoken to at the coffeeshop had also believed), the will of the people was being subverted, and these extraordinary events required an extraordinary response. The problem, my feed concluded, was the bad inputs these poor dupes had been fed. Inputs they had been fed due to a combination of deliberate opportunism from bad-faith political actors and gross negligence from our social media oligarchs.

My post needed a better ending, and the universe wrote it for me.


Is having bad inputs a solvable problem? My mind went to, of all things, a keynote I’d heard at a conference 30 days ago. The keynote had the delightful title Organizational epistemology or, how do we know stuff? and the speaker discussed the near impossibility of building consensus at scale. The solution, he argued, was designing data tools and organizations to bring people information that was relevant, timely, and trusted.

The thought wasn’t completely out of place, but what kind of person watches a coup unfold and thinks, this wouldn’t have happened if we had better data tools? If it wasn’t obvious to you by now, the answer I was going for was… Someone whose inputs are saturated with content about data tools and organizations.

A few years ago, I was designing a user tracking system for our company’s website and application. At the same time, I was taking a night class and did a project on San Francisco’s homelessness spending. My conclusion was that the city needed… A citywide tracking system for users of homelessness services.

Is there any better example that we can only see what we’ve been primed to see? Or only do what we’ve been primed to do?

Meesa propose that the Senate give, immediately, emergency powers to the Supreme Chancellor