SCRATCHPAD: 2024-07-18 Th: Biden Needs to Quit; & MAMLMs Powering a Redivision of Tech Surplus Towards Consumers; & MOAR...
A scratchpad...
A scratchpad…
Politics: I concur with this:
Dan Nexon: Would You Prefer a Side of Arsenic or Cyanide with Your Meal?: ‘Biden can win…. The dynamics of the election have plenty of time to change…. The fundamentals… are very favorable. However, the variable that points toward a loss, his approval rating, is all kinds of awful. I think that’s one reason why the “debate shock” proved particularly intense…. Biden and his advisors had been assuring everyone that everything was copacetic…. His performance in the campaign would bring people around. The opposite happened. Biden has done “okay” since then, he hasn’t been good…. Democratic officials have lost confidence that Biden can function as the party’s standard bearer. He seems incapable of consistently articulating a coherent, effective political message. Even when he does succeed, the message is now totally overshadowed by the “fitness to serve” questions. Democratic officials have simultaneously lost confidence in the White House itself as evidence accumulates that Biden’s staff was “managing him”… in some of the ways that we’d expect if Biden is in cognitive decline… <lawyersgunsmoneyblog.com/2024/07/would-…>
MAMLMs: It looks to me that tech oligopolists will now have to provide expensive LLC integration in order to preserve their customer base, without gaining (much) additional revenue thereby. Probably Google will gain a relative advantage because it does not have to pay the NVIDIA tax—but only a relative advantage: it will have less of a deduction from its profits. Perhaps Apple will gain a relative advantage as well—it may well avoid having to pay much NVIDIA tax, and it may well be able to transfer the electricity bill to its customers. LLC inference does not look like a tech business with revenue accompanied by effectively zero marginal cost to me. It looks, rather, like a constant-returns-to-scale high-cost low-revenue business:
John Quiggin: ‘Judging by the early proliferation, it seems that producing an adequate LLM isn't super-expensive. And It seems to me (not well informed) that they are already past the point of diminishing returns. As yet, most of them seem to be cheap or free and it seems quite possible they will stay that way, with profitability to be gained only by eating into the near-monopoly of advertising revenue now held by Alphabet and Meta… <substack.com/@jquiggin/note/c-61709785>
John Quiggin: ‘search with GPT is so much better than Google. Once (for example) Siri gives better answers to everything than Google does, it’s hard to see how Alphabet keeps making money…
Perhaps. Yes, if Apple figures out how to do on-device inference on its customers’ electricity and device-purchase dime, it is quite possible that Google will stop being a (very) profitable company.
But it might not work out that way. Apple may well fail to execute. And as long as Google maintains something like its current position in search, it will find a way to sell (at least some) ads, and it will not be paying the NVIDIA tax. It is, I think, very unlikely to be a more but it could become a not-much-less profitable company.
And as for the other highly-profitable tech oligopolists—they will wind up paying both the NVIDIA tax and the electricity bill. And for what additional revenue? Producer surplus looks to be going down, and consumer surplus looks (to me at least) to be going up.
Neofascism: This is what the Republican Base wants to hear, and the non-base hears—instead—that JOE BIDEN IS OLD. And people like me are told, over and over again, that they will not do it—that the point is to get some media events in which a few are deported, and to terrify others so much that they dare not ask for a raise: have a visible restoration of “public order” in which some transgressors are visibly punished behind barbed wire and by being pushed across the border. And it might “work” in that way:
Neil Stenhouse: ‘This is written like an exaggeration to make a point but it is actually just a clearly worded statement of the mainstream GOP position: Elias: ‘let’s all turn down the rhetoric and try to remain cool and clearheaded as we discuss the best way to build concentration camps for the 20 million people we have sworn to deport because they’re poisoning our national bloodstream… <bsky.app/profile/nstenhouse.bsky.social…>
Economics: This is nuts! When is the crash? The answer is that the crash—or, rather, the correction, or the rolling readjustment—will come soon after the vibe shifts from “those underweighting and shorting AI tech have a very strong fundamentalist case” to “only weirdos are fighting AI”. When that point comes, there is then no extra money to flip and drive AI-tech even higher:
Robert Armstrong: The only trade that matters: ‘Active fund managers are underperforming their benchmarks by even more than usual [in the first half of 2024]…. More than half of… returns in the S&P come from… Nvidia, Microsoft, Lilly, Meta and Amazon…. The five largest stocks in the Index (Microsoft, Apple, Nvidia, Alphabet, Amazon) account for 27 per cent of the total value of the index. For most active managers, having, say, 20 per cent of your fund in three stocks is a violation of risk limits; you are no longer diversified in any recognisable sense. But if you don’t have that much exposure to the biggest stocks in the index, you have almost certainly underperformed…. The options are (a) make a huge bet on big tech’s great run continuing, or (b) risk losing investor assets, followed by your job. This is not just a problem for stock pickers. Wall Street strategists are in the same bind…. Either you insist “the AI market” (as Unhedged has called it) is stable and sustainable, or you look like a dope. But this insistence requires you to turn your back on a lot of the standard principles of fundamental analysis and portfolio design… <https://www.ft.com/content/e05445ef-c6ae-4d42-9d1d-13e120ac3960>
Neofascism: That word “reänoint” bothers me, substantially:
Timothy Burke (2023): The News: Empires, Nations and Sovereignty: ‘The United States is going through another major wave of ethnonationalist struggle intended to reänoint the descendants of white Europeans as the only true ‘core’ of an immigrant nation… <timothyburke.substack.com/p/the-news-em…>
I remember landing in Boston in 1978 at the age of 18, and finding myself in a society in which when an Irish-American Catholic married an Italian-American Catholic people whispered about the “mixed marriage” and wondered if it would work over the long term. And I remember promptly being told by people who seemed to be of note and reputation that the “unmeltable ethnics” would lead a political revolution against the liberalism of the rootless cosmopolite WASP-Jewish alliance.
Remember the line from “The Good Shepherd”?
“Let me ask you something... we Italians, we got our families, and we got the church; the Irish, they have the homeland, Jews their tradition; even the ni——-s, they got their music. What about you people, Mr. Wilson, what do you have?” “The United States of America. The rest of you are just visiting.”
And since the days of William Jennings Bryan, if not before, on both the right and on the left, it has been very clear that the true “core” of the America people definitely does not include “Harvard professors, twisted-thinking intellectuals… the State Department [functionaires] burdened with Phi Beta Kappa keys and academic honors…”
“Reänoint”. No.
YIMBYism: IMHO, changing zoning is less than half the battle in defeating NIMBYism here in Berkeley. Streamlining the permitting process is much more important, even in the short run. “Looking at…” and “opportunities for input in 2024-2026” is still Polyanna:
Alfred Twu: ‘Downtown & North Berkeley buildings proposed & under construction. About 1,000 homes & 700+ dorm beds being built, another 5,000 homes & 1200-1500 beds planned. Almost 1 million square feet of classrooms & labs also…. Several new high-rise apartment buildings are planned, up to 28 floors. I remember when a 8-story building was a big deal, now it's unremarkable and many exist or are being built…. Like many college cities, Berkeley is building up. Besides demand for housing from more students, today's research universities have more researchers, and also spin off hundreds of startup companies with thousands of workers, all of who need places to live. With new state laws & local rezones, Berkeley is finally starting to catch up on having enough homes for the number of people working or studying here. We've still got a ways to go to reach jobs-housing balance, though finally we can see light at the end of the tunnel…. Berkeley City Council will be voting on the Middle Housing proposal to allow apartment buildings in the R-1 and R-2 zones that make up half the city…. Also Berkeley is finally looking at allowing larger buildings in the wealthier commercial districts of Elmwood, North Shattuck, and Solano Ave. This planning process just started and there will be opportunities for input 2024-2026… <x.com/alfred_twu/status/181034340654957…>
War: “When attacking equipment reveals itself, it can be destroyed and disabled very quickly”—that is an argument for why Ukraine will not lose the war quickly, not that Ukraine can win the war.
Ukraine has 1/4 the population and 1/3 the GDP per capita of Muscovy. Either the Russian régime somehow collapses, Ukraine loses, or Ukraine hangs on with the war fought with Ukrainian blood and NATO industry.
But what happens as the war drags on?
Reconquering the Donbas and Crimea for Ukraine will be very, very hard indeed, even with Ukraine willing to spill blood and NATO willing to build weapons:
Phillips P. O’Brien: The Last 6 Months Show Why Ukraine Can Win the War: ‘The fundamental advantages of defensive firepower (for now). The idea of a combined-arms advance of the type people were prophesying that Russia would do before February 24, 2022, is based on the ability to mass and supply forces of armored vehicles, supported by fixed wing aircraft and mobile artillery which would allow them to advance through and around enemy forces making rapid and deep movements. That form of war no longer exists for now. The range of defensive weapons, from accurate artillery, hand-helds, mines, and increasingly UAVs means that when attacking equipment reveals itself, it can be destroyed and disabled very quickly. We haven’t seen armored columns because whenever they try to be used in offensive roles they are located and hammered before they can advance… <https:// phillipspobrien.substack.com/p/weekend-update-86-the-last-6-months>
Abundance Agenda: Polarization was a thing created by Buchanan and Gingrich—and before them by Nixon, who shared their insight that plutocrat-loving Republicans would lose on lunchpail issues if they got the greater wealth inequality they sought unless they made politics all about Culture War. Nixon thus wanted to break the country in half, culturally, and come away with the bigger half. So did Buchanan and Gingrich. Their successors are happy to try to do the job with the smaller half plus stochastic terrorism. But a successful abundance agenda will not change that—unless it can develop a right wing with a principal allegiance to America rather than to plutocracy. From my perspective, the left-of-center Abundance Agenda is in very good shape. But the right-of-center? What is to be done there?:
Steven Teles & Rob Saldin: The rise of the abundance faction: ‘By working within a party, the new supply-siders can boost their cause—and pull us away from… polarization…. We can hope… the politically homeless… that used to be… in the Republican Party will reemerge… around their own version of Abundance rather than… exhausted… Reagan-era fusionism…. Sen. Todd Young of Indiana have already been leaders on many abundance-related issues. However… in the short term… the Democratic Party…. A faction… will require deliberate action in the form of organizing, mobilizing, and engaging at both the elite and mass levels…. An Abundance Agenda… would address… issues that, left unaddressed, will continue to be a drag on… our fellow citizens’ pursuit of happiness… [and] would enable a more dynamic, pluralistic, and representative future for American democracy… <http/s://hypertextjournal.substack.com/p/the-rise-of-the-abundance-faction>
Human Capital: This is the explanation that I find most convincing: being a conservative—as opposed to, say, being a believer in Whig measures implemented by Tory men—makes you think badly and stupidly.
I am gearing up to take another run at Steve Teles by pointing out that of the four Harvard conservatives that Teles names as liking for having positive VAR (well, three: I don’t think anyone even semiserious would say that about Harvey Mansfield, especially now that his anti-American ravings from the 2000s about how George W. Bush-as-Führer is above the law have become the law of the land, 6-3), all of them seem to have been made stupider by their conservatism vis-à-vis their Public Interest colleagues like Nathan Glazer and Daniel Patrick Moynihan who were not conservative.
When the young Samuel Huntington—a descendent of New England WASPs feels—compelled to write in The Soldier and the State that “the slaveowners were the only significant social group in the history of hte United States ever to be forcibly dispossessed of their property“, one has got to be brought up short. Do the slaves themselves not count as people “forcibly dispossessed of their property”? How about the Cherokee? The instinct that of these three groups only the slavemasters really exist means that there are a huge number of things that Samuel Huntington—even the “good” early Samuel Huntington—simply cannot see because he is a conservative:
Dan Drezner: ‘There’s a lengthy discourse about why so few conservatives are in the academy. It’s a multi-causal explanation, to be sure. But if it takes me, a political scientist, all of 10 minutes of reading to falsify the claims of a history Ph.D. who is also the president of the Heritage Foundation, then maybe one cause is that MAGA conservatives are really, really dumb?…
Yastreblyansky: ‘It’s always seemed plain to me—there’s not normally any real money in academia, from the conservative point of view, and they don’t see it as dignified. A conservative making a career in scholarship regards himself as a failure—a taker rather than a maker, feeding at somebody else’s trough. A real man is a participant in the creative destruction, not a mere observer. With exceptions in medicine, law, and maybe economics, but medicine is too hard, and the other departments are like “why would you do that? Sociology? Art history? Marine biology?”…
Economics: A 200,000 monthly payroll employment-growth number, but there are downward revisions to the past: That means that we now think we have only about 100,000 more more payroll jobs than we thought we had a month ago. That means the labor market has been slowing a bit more than we thought. That the labor market is putting even less pressure pushing prices upward then we thought. And, of course, all of these are very marginal—indeed, almost invisible—differences. The soft landing continues, without significant interruption:
Enda Curran: Here Are the Key Takeaways From US Jobs Report for June: ‘Nonfarm payrolls rose by 206,000 and job growth in the prior two months was revised down by 111,000…. The unemployment rate rose marginally to 4.1%, reflecting more people looking for work…. Average hourly earnings… annual increase… 3.9%… The median time it takes to find a job rose to 9.8 weeks, up from 8.9 weeks in May… <https://www.bloomberg.com/news/articles/2024-07-05/us-jobs-report-june-2024-key-takeaways-on-employment-payrolls>
Why, then, is monetary policy substantially restrictive? It is not that fiscal stimulus is making temporary r* higher than long-run r*—fiscal policy right now is a zero if not a drag on growth. Does anybody really believe that r* has risen substantially? John Williams certainly does not:
John C Williams: R-star—a global perspective: ‘The case for a sizable increase in r-star has yet to meet two important tests…. Plausible factors pushing up r-star… are likely to be global… a tension between the evidence from Europe that r-star is still very low and arguments in the United States that r-star is now closer to levels seen 20 years ago. Second, any increase in r-star must overcome the forces that have been pushing r-star down for decades… demographics and productivity growth…. Many of the explanations arguing for a higher r-star would likely show up in higher potential output growth. However, the HLW estimates of euro area and U.S. trend potential GDP growth in 2023 are nearly unchanged from their respective 2019 values… consistent with other estimates of potential GDP growth for the euro area and the United States… <https://www.bis.org/review/r240703f.htm>
The only caveat I have is that r* tracks expected future potential output growth rather than past. Thus MAMLM optimists have a case that r* has risern—but only a case, and that is the only case.
Neofascism: Tim Snyder has lessons for the present from the Nazi takeover back in 1933:
Tim Snyder: How to Stop Fascism: ‘(1) Voting matters. Hitler came to power after an election which enabled his appointment as head of government. It is much easier for fascists to begin from within than to begin from without…. (2) Coalitions are necessary….. The left has to hold together with the the center-left…. (3) Conservatives enabled…. They did not see Hitler and his Nazis as something different from themselves. They imagined, somehow, that Hitler would preserve the system rather revolutionize it. They were wrong…. (4) Big business should support democracy. In the Germany of the 1930s, business leaders were not necessarily enthusiastic about Hitler as a person. But they associated democracy with labor unions and wanted to break them…. (5) Citizens should not obey in advance. Much of fascism is a bluff…. American fascism… too is largely bluff, most of it digital…. Recalling history, we act in the present, for a future that can and will be much better… <snyder.substack.com/p/how-to-stop-fasci…>