David Watson 🥑

@EN_parker @mateosfo @dabluck @KelseyTuoc What do you think of “rents haven’t dropped in any city building new market-rate housing. Zero renters helped. The unit isn’t the primary vector of unaffordability, it’s merely the last component. Talk to a housing expert sometime”

@mateosfo @dabluck @KelseyTuoc @FifthRocket Here's a good one for this crew: A city upzones a neighborhood and sees a 15% increase in housing starts. After the new units are built, there has been no fall in rents. Does that mean the upzoning failed? Or can we say "this is good but doesn't solve the whole problem?"

Traffic deaths have not gone down in any city where Waymo is operating. Zero lives have been saved. That said, we will know in ~ 5-10 years if Waymo is better than the best human drivers (they only release data that compares them to the worst human drivers).

@mateosfo Another ! manifold.markets/DavidFWatson/w…

@FifthRocket Yes. Mathematically impossible, unless those cities do the actual things that reduce traffic deaths. Waymo will have 12,000 vehicles nationwide in 2030. It will still be a science experiment.

@dabluck @mateosfo @KelseyTuoc @EN_parker I mean, if it seems like we’ve reduced deaths in the lowest risk kids, that would be interesting…

@FifthRocket @mateosfo @KelseyTuoc @EN_parker I guess Matthew is saying if you inject only a few newborns and you focus on the ones who didn't need it, then how are you sure you saved lives?

@mateosfo @dabluck @KelseyTuoc @EN_parker Maybe more like vitamin k injection for newborns. A treatment applied to all to save a few. And we don’t measure that at the state level, why would you do that when we have more specific stats?

@dabluck @KelseyTuoc @FifthRocket @EN_parker Exactly right. She's proposing the equivalent of treating cancer patients by giving the treatment to people who don't have cancer. Total category error.

@mateosfo @EN_parker @KelseyTuoc I’m mostly just tired of this debate style. Yes I understand these are all related claims, but rather than debate their likelihood, it would be nice if we could agree on something measurable that might cause each of us to change viewpoints were it to happen.

@FifthRocket @EN_parker @KelseyTuoc The fact you think these are "four different goalposts" is less an observation about me and more a statement about your refusal to engage the topic sincerely. If I thought it would work, I'd literally draw you a picture. But seems your "interest" is mostly bad faith.

@mateosfo @EN_parker @KelseyTuoc You’ve pivoted through like four different goalposts for Waymo: aggregate city deaths, dangerous drivers, marketing claims, now Waymo doesn’t protect pedestrians from other cars I guess? Next it’ll be what, fleet turnover? Can you please make a single falsifiable prediction?

@EN_parker @KelseyTuoc @FifthRocket Yes exactly: Waymo can not solve the problem of humans crashing into Waymos. That is my entire point. Most human drivers never, ever crash and never, ever kill anyone. That has had literally no bearing on their risk of traffic death. Waymo is focusing on the wrong problem.

@KelseyTuoc @mateosfo @EN_parker I had this same argument with Matthew a few months ago. He basically won’t believe Waymo is improving safety till it’s probably at airline safety. Created this manifold market: manifold.markets/DavidFWatson/w…

@mateosfo @EN_parker They currently constitute about, what, ...0.006% of VMT in the US, mostly relatively safer ones? So they are probably displacing between 0.002% and 0.003% of the casualties we'd otherwise experience. They have started doing that. It just isn't showing up, obviously, in aggregates

@deanwball I appreciate a combined Hobbes/Corey title reference

@deanwball @ostalugo The Jones act blocks the US from installing the most cost effective offshore stuff, but I get the impression wind installations outside the US aren’t rising just due to subsidies. They must be providing value.

I'll restate: I have a principled opposition to wind subsidies, which have been a genuine boondoggle (for example: in the US it is common replace major, good-condition parts of extent wind turbines just to keep them subsidy-eligible). And were it not for subsidies, this technology would barely exist. So my principled opposition to wind is not literally to the physics of wind turning a turbine, but instead by way of the subsidies.

I have truly never understood how solar-maxis intend to deal with this reality; I expect there to be data centers 100 times this nameplate power draw in the nearish future. What you see below is 100mw. The response I usually get is "America has a lot of land," which is just bleak. Indeed, it turns *me* into a doomer, invoking as it does the notion of machines papering over our soil (which powers us) to power themselves. And it's not just data centers. In a world with electric freight trucks, a *truck stop* might require as much solar as you see pictured here, if not much more. A truck stop! Solar is fine; I do not have a principled opposition to it (which I do to eg wind). But solar's lack of energy density makes the solar-maximalist future a "loser premise," to borrow a phrase--at least it is a "loser premise" for human dignity. The good version of the future is of course a mix of many energy sources, but with a heavy bent toward fusion/fission and geothermal.

@Noahpinion @mattyglesias @KelseyTuoc I think the folks under discussion are the CEOs/leads. A survey of researchers can’t really tell you much about that. A survey of Bay Area Tech employees would say essentially none support Republican candidates.

I know @mattyglesias and @KelseyTuoc think that AI industry people are authentic doomers, but that's only true of SOME of them. Many were just dooming as a marketing pitch.

@DPearsonPHL @LeahLibresco One notable part of what Piper says here is that specifically Opus 4.7 was able to do this reliably and 4.6 was entirely incapable of doing it.

@LeahLibresco I asked Claude to try and assign editorials to authors, and it really struggled. Maybe it has gotten better since then, but I was actually a little surprised that it failed so badly.

It’s so over for Locke and Demosthenes “What if I try a college application essay I wrote 15 years ago, when my prose style was vastly worse and frankly embarrassing to reread? ‘Kelsey Piper,’ said Claude” theargumentmag.com/p/i-can-never-…

@TheZvi @Nition But… did the prompt say 6 fingers?

@Nition Opus generated a prompt based on the model card, and I fed it into Gemini and ChatGPT and chose the best output, which in this case was ChatGPT.

@mattyglesias Give them a break, in a few years the tree canopy will come in and it’ll be a bit more livable.

@Noahpinion If it’s not subscribers getting mad, presumably it’s not so bad? It’s not exactly a dire case of audience capture, but why worry?

Writing blog posts about architecture and urban design is very difficult because most relevant images are copyrighted and readers will get mad if I use AI images to make a point.

History is clear: Bombing civilian infrastructure -- does NOT trigger uprisings -- DOES trigger rage, nationalism, retaliation From WWII to today— This is bombing to lose, not bombing to win

@AndyMasley Nice! Yes, that’s basically my exact process as well. I made this for a friend youtu.be/HHIJhdHeSFE

@FifthRocket They're from Claude with some modifications

Suno just released a new model and I'm finding it to be a big shocking improvement where it's becoming very hard to detect the hints that it's AI music. Here's "I am actually scared of linear algebra"

@AndyCarnevale @mattyglesias

@FifthRocket @mattyglesias Never miss an opportunity to extol land value taxes!

@AndyCarnevale @mattyglesias (Pretty sure Yglesias is joking)

@mattyglesias Henry George proposed a land value tax on the *unimproved* value of land. This is land improvement and therefore wouldn't be taxed.