Another thing I've been thinking about is a retreat on something like "high-integrity AI x-risk comms" where people who care a lot about x-risk and care a lot about communicating it accurately to a broader audience can talk to each other.

I think this is a great idea that would serve an urgent need. I'd urge to you do it in the near future.

Agree with both the OP and Habryka's pitch. The Meetup Organizers Retreat hosted at Lighthaven in 2022 was a huge inflection point for my personal involvement with the community. 

My conflict got bumped to next week, so I'll make it after all. Have a large strawberry cheesecake to give for the cause.

• there's likely a kind of "position warfare" going on in policy circles to make sure your position is the one that's primed to win if the conditions for its enactment are suddenly met

I'm not sure how much "position warfare" happens for all but the most predictable events. After policy surprises there's usually a big fight to claim credit for predicting the crisis and a mandate for what to do about it. People and organizations certainly prepare for that (and I think our community should prepare more), but it's more by making predictions, finding allies, refining arguments, and writing plans.

• A shocking event led to the dominance of a political faction that previously had just been one of several competing factions, because that faction’s basic vibe (that we should make use of American hegemony, and that rogue states are a threat to national security) was roughly supported by the event.
• The response was substantially driven by elite judgements rather than popular judgement.

I think this is entirely correct. The Iraq War is one of the best examples of outside-the-Overton-Window policy change in recent memory. 

In my understanding, the key trigger for the "Milton Friedman Model of Policy Change" is the Policy Community being surprised. At its core, the Overton Window is a set of norms enforced by this community. In the wake of crisis those norms aren't enforced, so rather than shifting in some linear way, the window is temporarily suspended. Then, as Friedman said, "the actions that are taken depend on the ideas that are lying around." Thalidomide is another great example of when the policy change in the wake of a crisis has little to do with the trigger other than a particular faction winning the narrative fight.

I've been meaning to write more about this, would any particular angles be helpful?

I'm glad Jenn came to town and helped fix this policy proposal. I think her claim to 1-3% credit is an underestimate. I'm grateful for the first-hand account of how to fix policy that will not have the effects that the issuing body intends. Sometimes the best "technical correction" to a bad policy proposal is just start over.

The DC Policy Community is smaller than most people think. It's smaller than many people who are in it think. This miscalibration causes problems. There's a fundamental tension between performing your role to the best of your ability and taking on responsibility for the whole outcome. I think that tension, fiat justitia ruat caelum vs heroic responsibility, is key to understanding the bureaucratic soul. But that will need to be its own post someday.

Millions of people are employed in industries that export. The flaw in this policy proposal was easy to understand using only publicly-accessible data. Wikipedia gets you most of the way there. How is it possible that one small non-profit was the only organization to officially point out this basic flaw?

Jenn's explanation is great and entirely correct, but I'd like to highlight one of the drivers: something like the bystander effect. Balsa was not the only group to notice this problem. Everyone knew that this problem was easy to spot, knew that it must have been easy for the authors of the policy to spot it. But pointing it out has costs, risking the relationship if you're right, looking dumb if there's something you missed. Once the obvious error is pointed out in the record, even once, that should be enough to prevent the worst version of the rule from going into effect (by threat of an Administrative Procedures Act action, if nothing else). So if the pool of public commenters is large, and you're a prominent-but-not-overly-powerful group that has higher-priority problems on your plate, it's entirely rational to want someone else to bear the cost of pointing out the problem.

The problem comes if you're miscalibrated about how big the policy community is. If you expect an action to get thousands of well-informed public comments, it's probably safe to assume most such flaws will be pointed out. If the action only gets 586 comments in total, and most of those are focused on a more headline-grabbing aspect of the proposal, that assumption is no longer safe. 

There are hints of this in the record. The National Retail Federation's analysis implies there were back-channel communications. Their written testimony makes clear that they spotted the problem, yet chose not to mention it directly. "Surely someone else will point out this problem so we don't have to..."

This story has a moral: Do not assume what is obvious to you is obvious to the DC Policy Community. Even if you're right and the issue is obvious, you might still be the only person to speak up about the issue.

Not on my screen, not for the line breaks at least. See here:

https://thezvi.substack.com/p/balsa-update-springtime-in-dc

Though you're right, the footnotes aren't links in the substack version either. 

(Formatting Note: line breaks and footnotes got broken in the cross-post)

Note this event says June 27th. Was this intended? Thought I was RSVPing to one this coming Friday, June 20th 

It's certainly plausible that you're right, but I worry about this a lot more now after the supply chain disruptions from Covid and tarrifs. I worry that we'd have real cold-start bottlenecks that would take years to resolve, not weeks or months, in any scenario where we lose access to Chinese parts. Scenarios in which ocean shipping is substantially disrupted are even scarier in one sense, though China would probably be symmetricaly affected, or worse.

The best counter-argument to my worry, and biggest update I've had on this in recent years, is the success of the TSMC chip fab in Arizona. I predicted it would not go well. I'm delighted to have been wrong.