Bureaucrat. Penguin enthusiast.
JohnofCharleston
Karma: 209
Defense technology production is no longer about manufacturing. It’s culturally artisanal. In peer competition, the ability to scale is typically more decisive than artisinal quality (with the notable exception of the Manhattan project).
I think you’re right that American drones would likely be several times better, perhaps an order of magnitude better, than Chinese comparators. But we would have substantial bottlenecks on scaling production, could not simply resolve those bottlenecks with money, and even if we manage to scale appropriately would face huge cost disadvantages.
That’s a huge problem when the use case is swarm tactics. China could probably afford a strategy to saturate defenses with its drones, we probably could not, with our current mindset and processes.
Back at home now, this album is excellent at summoning the particular mood blending intensity, playfulness, and weltschmerz[1] that Lighthaven consistently instills in me.
Note: “You Have Not Been A Good User” seems to be missing, with 7 seconds of silence as a placeholder.
I didn’t listen closely to the lyrics two weeks ago, but I’m finding time now. “Friendly Fire” is particularly affecting. I don’t see a co-author listed, no obvious hits on the language. Was that you? How are you doing?
- ^
Claude’s helpful suggestion for “word for ennui + dread + hopelessness + obligation + duty”
- ^
Is Republican opposition to AI Regulation “lasting”? I notice there’s some movement this week on AI Whistleblower Protection from Congressional Republicans.
My read of the situation is that Republicans largely oppose whatever Biden supported, but weakly. That opposition can and has been overcome in several policy areas. Trump repealed the Biden AI Executive Order, and JD Vance gave a critical speech at the Paris Conference, but I wouldn’t assume their opposition is lasting or categorical.
I’d be very interested to read a post on this, though! If you have more thoughts on how sticky their opposition is, please write them up.
This marginal analysis is correct on its own terms, but I think is irrelevant to Jason’s point.
To escape American Federal or Californian regulation, a company would have to stop doing business in the relevant jurisdiction, not just shift marginal investment in new data centers. My understanding of Jason’s point is that the AI companies are willing to pay large costs already to stay where they are. Sure, there might be a tipping point where companies start leaving the Bay, which is a distinct risk of strong Californian regulation. But it seems implausible that any of the relevant companies would stop selling services to Americans.
Earning the Chance to Persuade
It’s not enough to just articulate a good idea and provide evidence that it’s good; you also have to do quite a bit of political advertising.
I’m glad Jason is explaining his theory of change in such detail. His points that governance ideas won’t spread automatically, and that we have to earn credibility for pitches to get traction, seem obvious in retrospect. But despite thinking about this topic quite a lot the past six months, I hadn’t fully incorporated this into my model.
In my one LW post to-date, I explained Milton Friedman’s theory of policy change, that “Only a crisis — actual or perceived — produces real change”. While I don’t think that’s the only way policy can significantly change, I believe it’s the main way that outside the Overton Window policy ideas become possible. To effectively use this theory, a community need only: make novel, accurate, and unpopular predictions about future political shocks, have a coherent model of the world that explains that shock, and at a crucial moment hand a binder to surprised and confused policymakers. If all goes well, they’ll be grateful for the suggestions and good policy ideas will promptly sail through Congress and rule-making agencies. That’s just three things, easy.
Coca-Cola famously spends massive amounts on advertising, but very little of it has an immediate, measurable impact. Coca-Cola doesn’t typically ask you to click on banner ads, to sign up for a subscription service, or to CALL NOW. Their advertising is trying to make sure that when you’re thirsty, later, you remember their brand, have positive associations, and will prefer their products over similar alternatives. This is basic Brand Awareness Advertising, and it works well when you can credibly predict that a decision will need to be made at some point, but not necessarily when, where, or the precise scenario in which it will be made.
Jason makes an excellent point, that should have been obvious, that when the next policy window opens, this community won’t be the only one claiming credit for the prediction. This community won’t be the only one offering a binder of policy fixes to confused and grateful policymakers. We need to do much more to prepare for that eventuality.
Brought my car to the office, hopefully there by 6.
I’m going and looking forward to it!
The Milton Friedman Model of Policy Change
Believe this time is wrong, should be Noon EST per the ACX post:
https://www.astralcodexten.com/p/spring-meetups-everywhere-2024
Washington DC Astral Codex Ten Meetup for Meetups Everywhere 2023
Washington DC ACX Mini-Meetups Everywhere—Spring 2023
Thank you for the update. I passed this listing on with my endorsement to several people in the DC community who applied and heard nothing back. Having the event fall through like this is discouraging.
In the future I’d urge you to share bad news sooner, it doesn’t get better with time.
As a counter-argument, Dominic Cummings’s argument that it is ~impossible to reform bureaucracies. The only plausible path to improvement is replacing moribund institutions with new startups, explicitly with different institutional cultures.
https://dominiccummings.substack.com/p/regime-change-2-a-plea-to-siliconNot endorsed, but worth grappling with.
Potentially relevant extra-credit reading, Scott Alexander’s Book Review of Seeing Like a State by James Scott:
https://slatestarcodex.com/2017/03/16/book-review-seeing-like-a-state/This perspective is foundational to my understanding of bureaucracies and other large institutions.
Washington DC Astral Codex Ten Meetup for Meetups Everywhere 2022
Something doesn’t add up with the Private Sector Vaccine Mandate section.
The claim is:
“US companies firing fully vaccinated people who don’t have the *right* US vaccines. Or people who even if unvaccinated have no way to get a US-approved vaccine in their own country.”…I don’t buy it. I simply do not believe American firms are firing ALL THEIR WORKERS in countries that do not have access to Pfizer, Moderna, or J&J. We would know about that. I don’t even believe that companies are firing significant fractions of their workforce in countries like the UK, who DO have theoretical access to the American vaccines, but mostly used others. The person you cite provides no evidence for that extraordinary claim, despite several people asking.
So what’s the real story? Is there a kernel of truth buried somewhere in there, like multinational corporations are requiring their staff IN THE US who were vaccinated overseas to get re-vaccinated with an American vaccine? I can easily believe that, but it’s a much less explosive claim. The specific claim is almost certainly not true.
Is anyone else confused about this?
Washington, DC – ACX Meetups Everywhere 2021
Be very careful about introducing an expectation to pay. Payment, even contingent future payment, will fundamentally change the nature of the endeavor. People are very motivated by social norms in most situations, but introducing money buys your way out of those norms. The classic treatment of this is is that daycare centers that fine parents for tardiness and making the staff stay late have much more tardiness than those who just disapprove.
See https://www.nytimes.com/2005/05/15/books/chapters/freakonomics.html
To the extent that this practice stays within a community, the community’s own norms are probably sufficient. Besides, many/most hosts would probably prefer guests donate to charity if they end up well-off. Maybe you have some sort of soft norm that guests offer to let their prior hosts direct some of their altruism, but be careful even with that. Everything matters on the margin, and that’s about the upper limit of incentive that won’t meaningfully displace the group norms.
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.