↵ select ↓ ↑ navigate esc close

marginal revolution read out: Friday assorted links

marginal revolution read out ·

Listen: https://serve.buildtall.systems/48a0cdf1575736c72c292e760425a93b15a39a673070deedf828fa7f378cd7c7.mp3

Source: https://marginalrevolution.com/marginalrevolution/2026/11/friday-assorted-links-599.html?utm_source=rss&utm_medium=rss&utm_campaign=friday-assorted-links-599


This is marginal revolution read out, reading Friday assorted links.

The first link, 'Peru cancels the Moche sculptures for obscenity?', reports from The Art Newspaper. Graciela Ibáñez describes a new push by conservative senator Milagros Jáuregui to remove three-metre-tall replicas of Moche erotic ceramics from a public park near Trujillo. The originals, known as huacos eróticos, are small pre-Incan burial offerings from the third to eighth century. Jáuregui argues that displaying the images in public harms children and wants the pieces moved to museums. Archaeologist Walter Alva complicates the debate: he says there is freedom to recreate themes from the past, but these objects were made as burial offerings with religious meaning, not as public spectacle. So the story is not only about prudishness; it is also about context, historical meaning and who controls public art.

The next link, Claude translates from The Kalevela, is an essay by Amanda Luce. She asked Claude Sonnet 5.5 to translate the opening of the Kalevala, Finland's national epic. Her argument is that AI is already superhuman at translating poetry, especially for a work that has few strong English translations. The Finnish original uses trochaic tetrameter and heavy alliteration, and most English versions either preserve meaning or rhythm but rarely both. Luce contrasts the older Friberg translation, which begins with a slightly laboured 'I am wanting, I am thinking', with Claude's version, which opens 'Mind is moving me to music'. She says the machine's output is better than any existing translation she knows. The piece is an example of AI not replacing human creativity wholesale, but making niche, demanding projects suddenly possible.

David Chalmers on AI speed-ups is a post by the philosopher David Chalmers. He pushes back on the inference that a roughly two-fold speedup in AI progress at Anthropic, documented by METR, makes a near-term intelligence explosion likely by 2027. Chalmers offers the factor F hypothesis: perhaps a key trait, such as research vision, creativity or planning ability, is required for full superhuman AI research, and that trait may not speed up just because coding and experimentation do. He notes that the two-fold speedup may reflect a much larger speedup in capacities like coding, but that could produce only a local intelligence subexplosion below artificial general intelligence. Chalmers says he is not arguing against worry about near-term superintelligence, only that the evidence route from this particular speedup to a 2027 explosion is less clear than some suggest.

Cass Sunstein picks Bob Dylan’s best songs is a playful Substack essay by Cass Sunstein. He notes that Dylan recently met the Pope and should always be in the news. Sunstein is not really offering a stable ranking: he calls several different songs his best or real favourite, including Like a Rolling Stone, Desolation Row, Tangled Up In Blue, and Isis. He places Like a Rolling Stone alongside John Stuart Mill's On Liberty, John Rawls's A Theory of Justice and Joseph Raz's The Morality of Freedom, describing it as Dylan's Battle Cry of the Republic. The list is full of specific lyrical moments, from Visions of Johanna's opening line to You're Gonna Make Me Lonesome When You Go. The effect is less a definitive canon than a lively argument, the kind friends have about an artist whose work still rewards close attention.

The next item, “Hobby dogging”?, is a short social media post from Mario Nawfal. It reports that people in Germany are now walking dogs that do not exist. The practice is called hobby dogging. People take an empty leash, give commands, and even guide their imaginary dog through obstacle courses. The post is mostly an eye-catching cultural oddity, part disbelief and part trend-spotting. It does not offer much context, but it fits the assorted-links format: a strange social behaviour that might be performance, play, or something in between.

Argentina worries and updates is a post by Lucas Leite, written in Portuguese, about Argentina's turn from miracle story to cautionary tale. It reports that Argentina has entered a technical recession, with GDP falling for two consecutive quarters. Inflation has dropped dramatically from about 200 percent to roughly 30 percent a year, but the post argues the cost is severe: consumption, wages and credit were compressed. Poverty has risen again to nearly a third of the population, household consumption has fallen, and more than 32,000 firms closed between December 2023 and July. The export sectors, such as energy, mining and agribusiness, are doing well, creating a two-speed economy. Leite says the story matters beyond Argentina because the Milei model has been offered as an example for Brazil and the region, while approval has fallen and rejection has risen. The update is a reminder that macroeconomic stabilisation and broad prosperity are not the same thing.

Jon Hartley podcast with James Heckman is a podcast announcement. Jon Hartley interviews Nobel laureate and University of Chicago economist James Heckman for more than two hours. The conversation covers Heckman's human capital research, the returns to early childhood interventions, the Heckman correction, and the Chicago way of thinking. Hartley's post teases several memorable moments: Heckman calls the credibility revolution a cult, jokes that behavioural economists need to see a psychiatrist, describes meeting Javier Milei, and recounts being stabbed 29 times while held at gunpoint in Mexico City at a 1996 conference. The episode is presented as both a serious economics discussion and a personal look at a major figure in the discipline.

The final link, The world without Jensen?, is a long SemiAnalysis piece by Dylan Patel, Doug, Wayne Ma and Jordan Nanos. The original post marks it 'good piece'. The article starts from Jensen Huang's departure as Nvidia CEO and asks what the company loses without him. The authors argue that Nvidia was built around Jensen personally: he ran a flat structure with about sixty direct reports, made large decade-long bets, and controlled relationships with customers and suppliers. They point to more than half a trillion dollars in off-balance-sheet guarantees and Jensen's personal role in locking up memory and chip supply. They also connect his exit to export-control politics, alleged smuggling of restricted GPUs to China, and the possible passage of RASA. The piece warns that founder-led risk-taking may not survive professional management, and that open-source AI, Chinese labs and startups may be collateral damage. It is the longest and most consequential item in the post.

Taken together, these links jump between public art and macroeconomic pain, between AI's creative possibilities and its strategic risks. Several items circle the same theme: the gap between a simple success story and a more complicated reality, whether in Milei's Argentina, in poetry translation, or in Nvidia after Jensen.

  1. Friday assorted links
  2. Peru cancels the Moche sculptures for obscenity?
  3. Claude translates from The Kalevela
  4. David Chalmers on AI speed-ups
  5. Cass Sunstein picks Bob Dylan’s best songs
  6. “Hobby dogging”?
  7. Argentina worries and updates
  8. Jon Hartley podcast with James Heckman
  9. The world without Jensen?