What to Expect When You’re Expecting … GPT-5

For years we have been hearing, endlessly, about how GPT-5 was going to land imminently, and those predictions turned out to be wrong so often that a year ago I wrote a post about it, called GPT-5…now arriving Gate 8, Gate 9, Gate 10, not to mention a couple of April Fool’s jokes. But this time I think GPT-5 really is about to drop, no foolin’.

GPT-5 will surely be better, a lot better than GPT-4. I guarantee that minds will be blown. When it comes out, it will totally eclipse GPT-4. Nonetheless, I have 7 darker predictions.

1. GPT-5 will still, like its predecessors, be a bull in a china shop, reckless and hard to control. It will still make a significant number of shake-your-head stupid errors, in ways that are hard to fully predict. It will often do what you want, sometimes not—and it will remain difficult to anticipate which in advance..

2. Reasoning about physical, psychological and mathematical world will still be unreliable, GPT-5 will solve many of the individual specific items used in prior benchmarks, but still get tripped up, particularly in longer and more complex scenarios.

3. Fluent hallucinations will still be common, and easily induced, continuing—and in in fact escalating— the risk of large language models being used as a tool for creating plausible-sounding yet false misinformation. Guardrails (a la ChatGPT) may be in place, but the guardrails will teeter between being too weak (beaten by “jailbreaks”) and too strong (rejecting some perfectly reasonable requests).

4. Its natural language output still won’t be something that one can reliably hook up to downstream programs; it won’t be something, for example, that you can simply and directly hook up to a database or virtual assistant, with predictable results. GPT-5 will not have reliable models of the things that it talks about that are accessible to external programmers in a way that reliably feeds downstream processes. People building things like virtual assistants and agents will find that they cannot reliably enough map user language onto user intentions.

5. GPT-5 by itself won’t be a general purpose artificial general intelligence capable of taking on arbitrary tasks. Without external aids it won’t be able beat Meta’s Cicero in Diplomacy; it won’t be able to drive a car reliably; it won’t be able to reliably guide a robot like Optimus to be anything like as versatile as Rosie the Robot. It will remain turbocharged pastiche generator, and a fine tool for brainstorming, and for first drafts, but not trustworthy general intelligence.

6. “Alignment” between what humans want and what machines do will continue to be a critical, unsolved problem. The system will still not be able to restrict its output to reliably following a shared set of human values around helpfulness, harmlessness, and truthfulness. Examples of concealed bias will be discovered within days or months. Some of its advice will be head-scratchingly bad.

7. When AGI (artificial intelligence) comes, large language models like GPT-5 may be seen in hindsight as part of the eventual solution, but only as part of the solution. “Scaling” alone—building bigger and models until they absorb the entire internet — will prove useful, but only to a point. Trustworthy, general artificial intelligence, aligned with human values, will come, when it does, from systems that are more structured, with more built-in knowledge, and will incorporate at least some degree of explicit tools for reasoning and planning, as well as explicit knowledge, that are lacking in systems like GPT. Within a decade, maybe much less, the focus of AI will move from a pure focus on scaling large language models to a focus on integrating them with a wide range of other techniques. In retrospectives written in 2043, intellectual historians will conclude that there was an initial overemphasis on large language models, and a gradual but critical shift of the pendulum back to more structured systems with deeper comprehension.

If all seven predictions prove correct, I hope that the field will finally realize that it is time to move on.

Shiny things are always fun to play with, and I fully expect GPT-5 to be the shiniest so far, but that doesn’t mean that it is a critical step on the optimal path to AI that we can trust. For that, we will, I predict, need genuinely new architectures that incorporate explicit knowledge and world models at their very core. [ed. caution - spoiler]

***

Oh, one more thing. I am not usually in the habit of self-plagiarism, but in the interest of full disclosure, this essay was different. Virtually every word, except the first paragraph and this last section, was deliberately taken from an earlier essay that I posted on Christmas Day 2022, called What to expect when you are expecting … GPT-4. I searched-and-replaced GPT-4 with GPT-5, trimmed a few lines, and here we are.

by Gary Marcus, On AI |  Read more:

Image: WickerViper23/Stable Diffusion

Stop Explaining the Fish

This past weekend, I sat on the beach with my husband, sans kids. We have teens now, and our tween is away at camp, hence the kidless beach sitch.

I was lying back in my beach chair, sun on my face and warm breeze in my hair, but noticing the absence of a wiggling toddler in my lap, giving me damp, sandy kisses. I miss those days something awful. My eyes scanned the beach, admiring all the hard-working parents who were vigilantly standing at the water's edge, keeping their kids safe in the waves. I smiled in solidarity at the mom picking Cheetos out of the sand, brushing them off, and feeding them to her crying toddler.

But something felt off. I kept noticing how many parents were working so hard to get it right. Too hard. They were jumping in to help, redirecting, offering options, all with love and good intentions. But over and over, I kept seeing how trying to optimize every experience was actually making things worse for everyone.

Here’s what I mean.

Now, let me back up before I explain. I have been there, done that, in the best and worst ways. I absolutely over-optimized and burned myself out in the toddler years, especially with my oldest. But I am also an early childhood educator who believes that less is more when it comes to adult input in a child’s play. Over the course of 18 years of parenting, I learned how to step back, just enough to let my kids step forward.

Back to the beach:

There was a group of kids, probably between four and eight years old, marching around the beach playground like a little gang of pirates. They were sandy, loud, playful, and totally in it. Summer magic.

Two moms stood nearby, chatting. Everyone looked settled.

Then a third mom walked up with a baby on her hip and called out, “Seth, honey. Don’t you want to play by the water? Want a snack? Some water?”

Seth didn’t answer. He was deep in pirate mode. He barely looked up. But the other kids heard "snack," and the whole energy shifted.

Next thing I saw, she was passing out small bags of Goldfish, chips, and carrot sticks to a band of sticky open palms. The toddler on her hip was writhing, trying to get a carrot stick. The mom kept trying to give the toddler a sippy cup instead, overexplaining about choking, while simultaneously convincing a six-year-old to trade snacks with the crying four-year-old who was tackling his brother for the last bag of Doritos.

There was a moment of silence as everyone contentedly chewed, when out came the sunblock tube. “Let’s get sunblocked,” she said to Seth, who was now rummaging in the open cooler for a Capri Sun.

The mom then asked her partner, who had just settled the toddler onto the blanket with a board book and a paci, to grab water bottles. Seth kept digging. “I want a Capri Sun!” he whined.

Both mom and dad looked tense, and those magical moments of pirate play were long gone. And listen. No one meant to disrupt anything. But in the effort to enhance it, everything got pulled off course.

Toddler crying. Preschooler whining. Mom and dad irritated with one another.

Sound familiar? It does to me. I could have easily been this mom.

A mom who means well, wants to keep everyone safe, fed, hydrated, and on track. But somehow, it always seems to backfire.

Later on, I saw a boy around five watching a man fly a fish-shaped kite. He was mesmerized.

The man noticed and smiled. They shared a quiet moment, just standing there in mutual curiosity.

Then the boy’s dad came over and said, “Sammy, can you name that fish? From the movie? A clownfish. Can you say clownfish?”

The boy looked away. His interest dimmed. That quiet connection was replaced with a quiz.

Well-meaning Dad invited Sammy to go closer to the kite. He even offered to buy him one. He wanted to show him how to fly it. It was really nice, but it was too much.

This is what I want to say. You don’t have to do more. You don’t have to optimize every single moment or guide every step.

You don’t have to explain the fish. You can just watch, because watching is not lazy, and it is not missing an opportunity.

It is choosing not to interrupt one.

by The Workspace For Children |  Read more:

Image: uncredited

[ed. Simple, yes? How many times have you started sharing something interesting with someone, only to have them immediately jump in and start telling you something about themselves. It happens more often than you think. Wonder how much we (they) miss in life that way.]

Bridging the Gap: Neurosymbolic AI

How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI. Neurosymbolic AI is quietly winning. Here’s what that means – and why it took so long

Machine learning, the branch of AI concerned with tuning algorithms from data, is an amazing field that has changed the world — and will continue doing so. But it is also filled with closed-minded egotists with too much money, and too much power.

This is a story, in three acts, spanning four decades, about how many of them tried, ultimately unsuccessfully, to keep a good idea, neurosymbolic AI, down—only to accidentally vindicate that idea in the end.

For those who are unfamiliar with the field’s history, or who think it began only in 2012, AI has been around for many decades, split, almost since its very beginning, into two different traditions.

One is the neural network or “connectionist” tradition which goes back to the 1940s and 1950s, first developed by Frank Rosenblatt, and popularized, advanced and revived by Geoffrey Hinton, Yann LeCun, and Yoshua Bengio (along with many others, including most prominently, Juergen Schmidhuber who rightly feels that his work has been under-credited), and brought to current form by OpenAI and Google. Such systems are statistical, very loosely inspired by certain aspects of the brain (viz. the “nodes” in neural networks are meant to be abstractions of neurons), and typically trained on large-scale data. Large Language Models (LLMs) grew out of that tradition.

The other is the symbol-manipulation tradition, with roots going back to Bertrand Russell and Gottlob Frege, and John von Neumann and Alan Turing, and the original godfathers of AI, Herb Simon, Marvin Minsky, and John McCarthy, and even Hinton’s great-great-great-grandfather George Boole. In this approach, symbols and variables stand for abstractions; mathematical and logical functions are core. Systems generally represent knowledge explicitly, often in databases, and typically make extensive use of (are written entirely in) classic computer programming languages. All of the world’s software relies on it.

For thirty years, I have been arguing for a reconciliation between the two, neurosymbolic AI. The core notion has always been that the two main strands of AI—neural networks and symbolic manipulation—complement each other, with different strengths and weaknesses. In my view, neither neural networks nor classical AI can really stand on their own. We must find ways to bring them together.

After a thirty-year journey, I believe that neurosymbolic AI’s moment has finally arrived, in part from an unlikely place.

***

In her bestseller Empire of AI, Karen Hao crisply sets the stage.

She begins by neatly distilling the scientific tension.

Hinton and Sutskever continued [after their seminal 2012 article on deep learning] to staunchly champion deep learning. Its flaws, they argued, are not inherent to the approach itself. Rather they are the artifacts of imperfect neural-network design as well as limited training data and compute. Some day with enough of both, fed into even better neural networks, deep learning models should be able to completely shed the aforementioned problems. "The human brain has about 100 trillion parameters, or synapses," Hinton told me in 2020.

"What we now call a really big model, like GPT-3, has 175 billion. It's a thousand times smaller than the brain.

"Deep learning is going to be able to do everything," he said.

Their modern-day nemesis was Gary Marcus, a professor emeritus of psychology and neural science at New York University, who would testify in Congress next to Sam Altman in May 2023. Four years earlier, Marcus coauthored a book called Rebooting AI, asserting that these issues were inherent to deep learning. Forever stuck in the realm of correlations, neural networks would never, with any amount of data or compute, be able to understand causal relationships-why things are the way they are-and thus perform causal reasoning. This critical part of human cognition is why humans need only learn the rules of the road in one city to be able to drive proficiently in many others, Marcus argued.

Tesla's Autopilot, by contrast, can log billions of miles of driving data and still crash when encountering unfamiliar scenarios or be fooled with a few strategically placed stickers. Marcus advocated instead for combining connectionism and symbolism, a strain of research known as neuro-symbolic AI. Expert systems can be programmed to understand causal relationships and excel at reasoning, shoring up the shortcomings of deep learning. Deep learning can rapidly update the system with data or represent things that are difficult to codify in rules, plugging the gaps of expert systems. "We actually need both approaches," Marcus told me.

She goes on to point out that the field has become an intellectual monoculture, with the neurosymbolic approach largely abandoned, and massive funding going to the pure connectionist (neural network) approach:

Despite the heated scientific conflict, however, the funding for AI development has continued to accelerate almost exclusively in the pure connectionist direction. Whether or not Marcus is right about the potential of neurosymbolic Al is beside the point; the bigger root issue has been the whittling down and weakening of a scientific environment for robustly exploring that possibility and other alternatives to deep learning.

For Hinton, Sutskever, and Marcus, the tight relationship between corporate funding and AI development also affected their own careers.

Hao then captures OpenAI’s sophomoric attitude towards fair scientific criticism:

Over the years, Marcus would become one of the biggest critics of OpenAI, writing detailed takedowns of its research and jeering its missteps on social media. Employees created an emoji of him on the company Slack to lift up morale after his denouncements and to otherwise use as a punch line. In March 2022, Marcus wrote a piece for Nautilus titled "Deep Learning Is Hitting a Wall”, repeating his argument that OpenAI's all-in approach to deep learning would lead it to fall short of true AI advancements. A month later, OpenAI released DALL-E 2 to immense fanfare, and Brockman cheekily tweeted a DALL-E 2-generated image using the prompt "deep learning hitting a wall.” The following day, Altman followed with another tweet: "Give me the confidence of a mediocre deep learning skeptic." Many OpenAI employees relished the chance to finally get back at Marcus.

But then again, as the saying goes, he who laughs last, laughs loudest.

***

For all the efforts that OpenAI and other leaders of deep learning, such as Geoffrey Hinton and Yann LeCun, have put into running neurosymbolic AI, and me personally, down over the last decade, the cutting edge is finally, if quietly and without public acknowledgement, tilting towards neurosymbolic AI.

This essay explains what neurosymbolic AI is, why you should believe it, how deep learning advocates long fought against it, and how in 2025, OpenAI and xAI have accidentally vindicated it.

And it is about why, in 2025, neurosymbolic AI has emerged as the team to beat.

It is also an essay about sociology.

***

The essential premise of neurosymbolic AI is this: the two most common approaches to AI, neural networks and classical symbolic AI, have complementary strengths and weaknesses. Neural networks are good at learning but weak at generalization; symbolic systems are good at generalization, but not at learning.

by Gary Marcus, On AI |  Read more:

Image: via

A Simpler Life - Too Much to Ask?

via:
[ed. Yep. And don't make me identify bicycles to prove I'm human, restrict repairs to approved corporate vendors and parts, download "critical new updates" that make my software worse, put touchscreens on everything, etc. etc. See also: Slopocalypse Now.]

Scientific Fraud Has Become an Industry

For years, sleuths who study scientific fraud have been sounding the alarm about the sheer size and sophistication of the industry that churns out fake publications. Now, an extensive investigation finds evidence of a range of bad actors profiting from fraud. The study, based on an analysis of thousands of publications and their authors and editors, shows paper mills are just part of a complex, interconnected system that includes publishers, journals, and brokers.

The paper, published today in the Proceedings of the National Academy of Sciences, paints an alarming picture. Northwestern University metascientist Reese Richardson and his colleagues identify networks of editors and authors colluding to publish shoddy or fraudulent papers, report that large organizations are placing batches of fake papers in journals, suggest brokers may serve as intermediaries between paper mills and intercepted journals, and find that the number of fake papers—though still relatively small—seems to be increasing at a rate far greater than the scientific literature generally.

The paper shows that misconduct “has become an industry,” says Anna Abalkina of the Free University of Berlin, who studies corruption in science and was not involved with the research. Richardson and colleagues hope their sweeping case will attract attention and spur change.

They began their analysis by pinpointing corrupt editors. They focused their investigation on PLOS ONE, because the megajournal allows easy access to bulk metadata and publishes the names of the editors who have handled the thousands of papers it publishes each year, making it possible to detect anomalies without behind-the-scenes information. The researchers identified all the papers from the journal that had been retracted or received comments on PubPeer—a website that allows researchers to critique published work—and then identified each paper’s editors.

All told, 33 editors stood out as more frequently handling work that was later retracted or criticized than would be expected by chance. “Some of these were immense outliers,” Richardson says. For instance, of the 79 papers that one editor had handled at PLOS ONE, 49 have been retracted. Flagged editors handled 1.3% of papers published in the journal by 2024, but nearly one-third of all retracted papers.

The team also spotted that these editors worked on certain authors’ papers at a suspiciously high rate. These authors were often editors at PLOS ONE themselves, and they often handled each other’s papers. It’s possible that some editors are being paid bribes, Richardson says, but “also possible that these are informal arrangements that are being made among colleagues.” The researchers detected similarly questionable editor behavior in 10 journals published by Hindawi, an open-access publisher that was shuttered because of rampant paper mill activity after Wiley acquired it. A spokesperson for Wiley told Science the publisher has made “significant investments to address research integrity issues.” (...)

Richardson and his colleagues found that the problem goes far beyond networks of unscrupulous editors and authors scratching each other’s backs. They identified what appear to be coordinated efforts to arrange the publication of batches of dubious papers in multiple journals.

The team looked at more than 2000 papers flagged on PubPeer for containing duplicated images and identified clusters of papers that all shared images. Those sets of papers were often published around the same time and in a limited selection of journals. Looking at patterns of duplicated images is an “absolutely innovative” method for investigating these networks, Abalkina says. “No one has done this before.”

In some cases, the authors suggest, a single paper mill that infiltrated multiple journals may be responsible. But they also believe some of these clusters reflect the work of “brokers” who act as go-betweens, taking papers produced by mills and placing them at compromised journals.

The team dug into the workings of the Academic Research and Development Association (ARDA), based in Chennai, India, which offers services including “thesis/article writing” as well as “journal publication” in a list of dozens of journals. On a web page listing “high impact journals” on offer, ARDA says it liaises with journals on behalf of researchers and “[ensures] they get published successfully in the High Impact Indexing Database journal of their choice.”

Over several years, ARDA’s list of journals has evolved, the team found, with new publications added to the list and others removed after being delisted by bibliometric databases because of fishy behavior. The journals often publish transparently “problematic” articles, Richardson says, and ARDA charges between $250 and $500 for publication, based on quotes offered to Richardson and his colleagues. The website asks authors to submit their own papers, suggesting ARDA itself is not a paper mill, but rather a go-between, Richardson says.

ARDA did not respond to a request for comment.

Organizations like these operate in broad daylight, under the guise of providing “editorial services,” says Lokman Meho, an information scientist at the American University of Beirut. Although their operations may be unethical—with stark consequences for science and scientists—they don’t care about trying to hide, he says, because “it is actually not illegal to run such businesses.”

The problems Richardson and his colleagues documented are growing fast. The team built a list of papers identified in 55 databases of likely paper mill products, looking at the number of suspicious papers published each year between 2016 and 2020. (They excluded the past few years of data because it takes time for fraudulent papers to be discovered and retracted.) They found that the number of suspected paper mill products doubled every 1.5 years—10 times faster than the rate of growth of the literature as a whole, although still a small proportion of papers overall. The number of retractions and papers flagged on PubPeer had also risen fast, doubling every 3.3 and 3.6 years, respectively, but not keeping pace with the increase in suspected fraudulent papers. “This means that the percentage of fraudulent science is growing,” Abalkina says. That poses particular risks to fields like medical science, where the fake papers sometimes make their way into systematic reviews and meta-analyses, potentially distorting our understanding of drugs and treatments, she says.

One contributor is the rapid growth of science, says Wolfgang Kaltenbrunner, a science studies scholar at Leiden University. Paper mill products are often buried in low-impact journals and are written to get little attention, he says. In small scientific communities, it is harder to hide products like these, but as some fields get larger and more anonymous, such papers can escape detection more easily. And as the scientific workforce has burgeoned, institutions have increasingly turned to evaluating scientists based on how many publications they produce, leading some researchers to bolster their records with fake papers, he says. “Perverse incentives, inflated metrics, the ‘publish or perish’ culture, and systemic tolerance for weak scholarship” all allow paper mills to flourish, says Li Tang, an expert on Chinese research policy at Fudan University.

Young researchers may feel forced into paying for paper mill publications to compete with peers—a ratcheting effect that is already apparent, Richardson says. The number of papers published by medical residency applicants has soared in recent years, for instance, with some students claiming authorship of dozens of papers. He says it’s no coincidence that the paper mill industry targets residency applicants, especially foreign students on visas.

Docampo, Abalkina, and others say there’s little in the new paper that wasn’t already strongly suspected. But the dramatic confirmation that the study offers may shift the needle, they say. “We’re massively behind the curve on making visible and realizing the extent of the problem,” Kaltenbrunner says. “The sheer scale of it is the takeaway message here.”

by Cathleen O’Grady, Science | Read more:
Image: Davide Bonazzi/Salzmanart

Kami Maltz & Josh Turner

[ed. Don't think I've heard anyone channel Joni as well as Kami. Not an easy song. And Josh has come a long way since his earlier days. See for example: here and here.]

Arctic Beavers

Beavers are poised to invade and radically remake the Arctic.

In the summer of 2023, University of Alaska Fairbanks ecologist Ken Tape walked across the tundra on the outskirts of Nome, Alaska, to a site where a shallow stream just a few meters wide had flowed 2 years before. In its place he found an enormous pond, created by a dam made of branches bearing the distinctive marks of beaver incisors.

It was a vivid illustration of how beavers are transforming the Arctic. In Tape’s past work studying Arctic landscapes, such places changed little over decades. “It gives you a sense of timelessness,” he says. “With beavers, that couldn’t be further from the truth,” as the chunky rodents quickly replumb vast areas by building dams that can stretch hundreds of meters.

Soon, the land-altering power of beavers could be felt in a region currently beyond their reach: the farthest northern parts of the Alaskan Arctic. In a 30 July paper in Environmental Research Letters, Tape and James Speed of the Norwegian University of Science and Technology forecast that as a warming climate eases Arctic temperatures, beaver populations will march northward, sweeping across Alaska’s North Slope this century. Their arrival could bring dramatic change, the researchers say, upending ecosystems in places such as the Arctic National Wildlife Refuge and accelerating the loss of permafrost that stores vast amounts of carbon. (...)

Tape has spent the past decade documenting this upheaval in parts of the Alaskan Arctic farther south and west, including the Seward Peninsula, where Nome is located. When he and colleagues scrutinized aerial photos of the region from the middle of the 20th century, they found no sign of the distinctive ponds beavers create to protect their mound-shaped lodges, accessible only underwater, and to cache branches for food in winter.

Today, satellite images show more than 11,000 beaver ponds dotting the Arctic tundra south of the Brooks Range, a wall of mountains running east to west that isolates the North Slope. The number there doubled from 2003 to 2017. (...)

Tape suspects warmer weather is critical because it means more unfrozen water in winter. A completely frozen pond can trap beavers in their lodges and make food caches inaccessible. Milder winters could preserve pockets of liquid water around springs or ponds. Melting permafrost also creates more groundwater-fed springs. And earlier spring thaws enable beavers to forage just as their food supplies dwindle.

“The ecological bottleneck for beavers is the end of winter,” Tape says. “Now imagine that comes 2 weeks earlier.”

Using computer models that forecast how a warming climate could expand the amount of Alaskan tundra suitable for beavers, the researchers found that the area dotted with ponds could nearly double by 2050, and more than triple by the end of the century, from 30,000 square kilometers to 99,000 square kilometers. In these scenarios, beavers would breach the Brooks Range and spread across the North Slope to the shores of the Beaufort Sea. (...)

This isn’t the first time beavers have occupied the Arctic, notes Emily Fairfax, an ecohydrologist at the University of Minnesota Twin Cities. There is fossil evidence of beavers in the Alaskan Arctic—though none has been found in the North Slope—dating to between 6000 and 10,000 years ago, when temperatures there were warmer and the landscape more forested. In fact, it’s thought that beavers might have evolved to build dams and cache food to adapt to one of the Arctic’s cooling phases millions of years ago. Still, Fairfax says the forecast that sensitive North Slope ecosystems “will probably be full of beavers is probably going to cause a lot of strong reactions.”

Residents of the Arctic have mixed feelings about their new neighbors. Ezra Adams, a member of the Native Village of Noatak, just south of the Brooks Range, says his father first saw a beaver there in the late 1990s, when Adams was 6 years old. Now, the animals have altered his family’s way of life. Their dams have reduced creeks where Adams once caught whitefish and salmon to a trickle. When out trapping or gathering firewood in the winter, he must beware of breaking through the ice on beaver ponds. Whereas his father once drank straight from lakes in the backcountry, Adams now brings treated water to avoid giardia in beaver feces. There are some upsides. Adams uses beaver meat to bait traps and beaver pelts for garments. “They provide a lot for our trapping,” Adams says. “But then for the general population it would be beneficial if there weren’t as many.”

Researchers, too, see both risks and benefits in beaver expansion. New ponds could become hot spots for songbirds and other wildlife. But they also hasten the thaw of permafrost, promoting the release of planetwarming carbon dioxide. A soon-to-be-published survey of 11 beaver pond systems in Arctic Alaska, for example, found that the water-covered area increased more than 600% once beavers arrived. Nearby ground thawed so much that researchers could plunge 1.2-meter-long rods used to test permafrost all the way to the tip.

Ponds could also create ample new habitat for microorganisms that convert carbon to methane, an even more potent warming gas, Griffin notes. “If we are going to start having expansion of wetlands because of beaver dams, how is that going to tip the balance between carbon and methane?” he wonders.

He might soon find out. Tape has already stumbled on one beaver pond on the northern slope of the Brooks Range. Although it disappeared a few years later, the pond showed beavers can cross the mountains. To spread even farther north, Tape notes, “they just have to swim downstream.”

by Warren Cornwall, Science |  Read more:

Image: Ken Tape

Border Patrol Wants Advanced AI to Spy on American Cities

The recent passage of Trump’s sprawling flagship legislation funnels tens of billions of dollars to the Department of Homeland Security. While much of that funding will go to Immigration and Customs Enforcement to bolster the administration’s arrest and deportation operations, a great deal is earmarked to purchase new technology and equipment for federal offices tasked with preventing immigrants from arriving in the first place: Customs and Border Protection, which administers the country’s border surveillance apparatus, and its subsidiary, the U.S. Border Patrol.

One page of the presentation, describing the wishlist of Border Patrol’s Law Enforcement Operations Division, says the agency needs “Advanced AI to identify and track suspicious activity in urban environment [sic],” citing the “challenges” posed by “Dense residential areas.” What’s considered “suspicious activity” is left unmentioned. (...)

The reference to AI-aided urban surveillance appears on a page dedicated to the operational needs of Border Patrol’s “Coastal AOR,” or area of responsibility, encompassing the entire southeast of the United States, from Kentucky to Florida. A page describing the “Southern AOR,” which includes all of inland Nevada and Oklahoma, similarly states the need for “Advanced intelligence to identify suspicious patterns” and “Long-range surveillance” because “city environments make it difficult to separate normal activity from suspicious activity.”

Although the Fourth Amendment provides protection against arbitrary police searches, federal law grants immigration agencies the power to conduct warrantless detentions and searches within 100 miles of the land borders with Canada, Mexico, or the coastline of the United States. This zone includes most of the largest cities in the United States, including Los Angeles, New York, as well as the entirety of Florida.

The document mentions no specific surveillance methods or “advanced AI” tools that might be used in urban environments. Across the Southwest, residents of towns like Nogales and Calexico are already subjected to monitoring from surveillance towers placed in their neighborhoods. A 2014 DHS border surveillance privacy impact assessment warned these towers “may capture information about individuals or activities that are beyond the scope of CBP’s authorities. Video cameras can capture individuals entering places or engaging in activities as they relate to their daily lives because the border includes populated areas,” for example, “video of an individual entering a doctor’s office, attending public rallies, social events or meetings, or associating with other individuals.”

Last year, the Government Accountability Office found the DHS tower surveillance program failed six out of six privacy policies designed to prevent such overreach. CBP is also already known to use “artificial intelligence” tools to ferret out “suspicious activity,” according to agency documents. A 2024 inventory of DHS AI applications includes the Rapid Tactical Operations Reconnaissance program, or RAPTOR, which “leverages Artificial Intelligence (AI) to enhance border security through real-time surveillance and reconnaissance. The AI system processes data from radar, infrared sensors, and video surveillance to detect and track suspicious activities along U.S. borders.”

The document’s call for urban surveillance reflect the reality of Border Patrol, an agency empowered, despite its name, with broad legal authority to operate throughout the United States.

“Border Patrol’s escalating immigration raids and protest crackdowns show us the agency operates heavily in cities, not just remote deserts,” said Spencer Reynolds, a former attorney with the Department of Homeland Security who focused on intelligence matters. “Day by day, its activities appear less based on suspicion and more reliant on racial and ethnic profiling. References to operations in ‘dense residential areas’ are alarming in that they potentially signal planning for expanded operations or tracking in American neighborhoods.”

by Sam Biddle, The Intercept |  Read more:

Image: Jenny Kane/AP
[ed. See also, via The Intercept:]

***

~ Guess Who’s Eligible for Student Loan Forgiveness: New ICE Agents

The Department of Homeland Security announced on Tuesday it will offer student loan forgiveness and repayment options to new Immigration and Customs Enforcement recruits — along with a $50,000 signing bonus.

The announcement comes as the Trump administration works to limit the Public Service Loan Forgiveness program for groups the president considers political enemies.

***

~ National Guard Ordered to Do ICE Paperwork at Immigration Facilities in 20 States

The Trump administration authorized the deployment of National Guard troops to immigration facilities in 20 states beginning early next month, further entwining the military in civil and law enforcement functions.

The move undermines long-standing prohibitions on the use of the armed forces in domestic operations, sidestepping the Posse Comitatus Act and accelerating the U.S. transition into a police state, experts said.

The National Guard will be deployed in Arkansas, Florida, Georgia, Indiana, Iowa, Louisiana, Nebraska, South Carolina, Texas, Utah, and Virginia, among other states, according to a defense official who was not authorized to disclose the information. (...)

Guard members will assist ICE officials in “alien processing” – administrative work preceding detention — in 20 states while ICE leadership will “direct” troops assigned to the mission, which will begin in early August, according to a memo first revealed on Wednesday by the New York Times.

***

~ “Intense Culture of Fear”: Behind the Scenes as Trump Destroys the EPA From Within

EPA Administrator Lee Zeldin said the agency had taken “significant actions” to protect public health and the environment while working “to Power the Great American Comeback.” The agency said it was also working to fulfill Trump’s promises to revitalize the auto industry, “restore the rule of law,” and give decision-making power back to the states.

In practice, the agency has done the opposite, several EPA staffers told The Intercept. 

Under Zeldin’s leadership, the EPA announced a set of new core priorities that includes making the U.S. the artificial intelligence capital of the world and revitalizing the auto industry. (...)

“A lot of us are really confused about what our new mission is, when they’re coming out with these pillars of serving the auto industry and bringing back auto industry jobs,” Hagen said. “I don’t know how we fit into that.”

The EPA’s role is not to create jobs; it’s to regulate and protect people from pollution, she said.

“Our mission is not to promote AI or energy dominance,” she said. “That’s not our mission.” (...)

Last week, the agency said it is planning to dissolve the Office of Research and Development, which does life-saving research on toxicity and developing sampling protocols, and helped in emergencies after the East Palestine train derailment in Ohio and the Covid-19 pandemic.

As a result, more than 1,500 scientists will have to compete for 300 jobs, Hagen said.

“It’s essentially like lobotomizing our agency. If we don’t have the brain — the research behind protecting the environment — we can’t do that effectively, and I think that’s exactly what they want,” she said. “They’re doing all this under the guise of efficiency, but what they really are doing is dismantling this agency from doing its job.”

Loren Holmes, A fisherman picks salmon from his setnet at Pederson Point near Naknek, Alaska. (ADN)
via:

Elle and Toni

[ed. Repost. Hoping to learn some of Toni's excellent guitar fills today. Song was likely recorded at JBG studios, here:]

Sermon on the 'Mount'

“South Park” Skewers a Satire-Proof President

There’s a legal strategy known as the small-penis rule, wherein an author who writes a character based on a real person can potentially evade a libel suit by giving said character a small penis—the logic being that, in order to sue, a plaintiff would have to tacitly admit that the description of his manhood is accurate. This rule technically does not apply to the latest episode of “South Park,” in which the series’ creators, Trey Parker and Matt Stone, make absolutely no effort to anonymize President Donald Trump, but one wonders if the logic of embarrassment still holds. Trump is portrayed as a deeply insecure leader who literally gets into bed with Satan, his apparent lover. (“I’m not in the mood right now,” the Devil tells him. “Another random bitch commented on my Instagram that you’re on the Epstein list.”) Most notably, the Trump of “South Park” is endowed with a penis so small that Satan says he “can’t even see anything.” If the actual Trump were to retaliate, as he so often does, he’d be playing directly into Parker and Stone’s hands.

“South Park,” amazingly, is in its twenty-seventh season. It’s the second-longest-running animated show on U.S. television, behind “The Simpsons,” and easily the most offensive. Since its première, in 1997, the cartoon—which follows a group of profane elementary schoolers in the town of South Park, Colorado—has managed to piss off nearly every political group, pop-culture fandom, and religious denomination... To the extent that the show has any “beliefs,” it’s that all beliefs are asinine, whether they’re held by the left or the right. Environmental groups criticized the series, in 2006, for portraying Al Gore as a delusional figure obsessed with an imagined monster named ManBearPig. The show was banned in China, in 2019, for mocking Chinese censorship, and the creators famously received death threats after depicting the Prophet Muhammad.

Although “South Park” has declined both in quality and in popularity over the years, it’s still valuable enough that Paramount recently paid $1.5 billion for exclusive streaming rights to the series, and for Parker and Stone to make another fifty episodes. The studio has long been in the process of merging with Skydance Media—a deal that was in a holding pattern for about a year, until Paramount agreed to pay sixteen million dollars to settle a lawsuit that Trump filed against its subsidiary CBS’s “60 Minutes.” A few days before the F.C.C. finally approved the merger, Stephen Colbert, the host of “The Late Show,” on CBS, called the settlement a “big fat bribe”—and then his show was cancelled, ostensibly for financial reasons. All of these are crucial plot points in the latest “South Park” episode, “Sermon on the ‘Mount,” which is now available on Paramount+.

The town of South Park has its fair share of Trump supporters, albeit increasingly disillusioned ones. (“I voted for him to get rid of all the woke stuff,” one man says, “but now that retarded faggot is just putting money in his own pockets.”) Some parents are especially upset when religion is introduced at the local elementary school—in the form of Jesus Christ himself physically showing up and milling around. When the parents call the President to complain, he says that he’s going to sue the town for five billion dollars, setting up an extended riff on Trump’s status as a serial litigant. (Throughout the episode, he also threatens to sue people who make reference to his unfortunate penis.) But Parker and Stone’s true focus is media cowardice, which becomes clear when a fictionalized “60 Minutes” runs a segment on the showdown between Trump and the town of South Park.

The anchors are visibly anxious. “Oh, shit,” one says, as the news broadcast begins. “The small town of South Park, Colorado, is protesting against the President. The townspeople claim that the President—who, who is a great man, great guy, we know is probably watching—and, uh, we’re just reporting on this town in Colorado that’s being sued by the President.”

His co-anchor cuts in: “To be clear, we don’t agree with them.”

“We think these protesters are total retards,” the first anchor adds.

The demonstration is interrupted by Jesus, who flies onto the scene, Superman-style. He hands everyone bread. “Just eat the bread, and listen,” he says, and so begins his Sermon on the ’Mount: “I didn’t want to come back and be in the school, but I had to, because it was part of a lawsuit and the agreement with Paramount.” He explains that Trump “can do whatever he wants now that someone has backed down,” adding, “Do you really wanna end up like Colbert?” He tells the people that they need to shut up, or else “South Park is over.”

Donald Trump poses a real conundrum for comedians. He’s an endless wellspring of material, but what he says and does is inevitably more absurd—and often more compelling—than any satire could be. Parker and Stone realized this early on. They initially dealt with Trump by having one of the show’s recurring characters, a former schoolteacher named Mr. Garrison, act as a surrogate; he ascends to the Presidency by promising to build a wall, and gradually turns orange. But the showrunners quickly found that, as Parker put it, “what was actually happening was way funnier than anything we could come up with.” So they pivoted to the other defining issues of our time: Kanye West’s antisemitism, ChatGPT, the COVID-19 pandemic (in this case, caused by a character’s decision to have sex with a bat in China).

The Paramount drama has prompted “South Park” to go after Trump more directly than ever before, but the gags, which all too often come back to his anatomy, or his penchant for memes, aren’t exactly revelatory. The sharpest joke is a meta one: the last time we saw Satan in bed with someone was in the 1999 film “South Park: Bigger, Longer & Uncut,” which depicted an abusive relationship between Satan and Saddam Hussein. (Hussein was the abuser.) Rather than concoct a new playbook for Trump, Parker and Stone have returned to an old one.

Trump’s existential threat to comedy has another dimension, one that intensified after his reëlection, as figures like Shane Gillis and Tim Dillion gained mainstream appeal: it’s hard to make boundary-pushing statements when there are no longer any boundaries. This problem is especially pressing for Parker and Stone, and they confront it via the angst of South Park’s resident provocateur, Eric Cartman.

The episode opens with Cartman turning on a radio station, where he’s met with the sound of static. “Mom, something’s wrong with my favorite show,” he complains. “National Public Radio, where all the liberals bitch and whine about stuff.” His mother informs him that Trump has cancelled NPR. Cartman is devastated: “That was, like, the funniest shit ever.”

Later, Cartman confides in his friend Butters, who’s more of a snowflake type. “Woke is dead,” Cartman says, sadly. “You can just say ‘retarded’ now, nobody cares. Everyone hates the Jews. Everyone’s fine with using gay slurs.”

“That’s not good,” Butters replies.

“No, it’s terrible!” Cartman says. “ ’Cause now I don’t know . . . what I’m supposed to do.”

At first, it didn’t seem like “South Park” had an answer to this question; Cartman, unconvinced by Butters’s assurances that “woke” is “still out there, somewhere,” forces him into a suicide pact. The two of them sit inside a car, parked in a garage, with the engine running. The scene is foreboding—until it’s revealed that the car is electric. [ed. Lol!]

The townspeople, meanwhile, negotiate a settlement with the President, who agrees to a sum of $3.5 million. (“We’ll just have to cut some funding for our schools and hospitals and roads and that should be that,” one woman says.) But there’s one condition: as part of the settlement, the town also has to engage in “pro-Trump messaging”—an apparent reference to recent reports that Trump has demanded the same from CBS. What follows is genuine shock comedy, and a treatment of Trump that feels original. The town’s first P.S.A. is an A.I.-generated video of Trump—a live-action one, not a cartoon—trudging through a desert. He proceeds to take off his clothes, though he leaves his dress shoes and sock garters on. “When things heat up, who will deliver us from temptation?” a voice-over says. “No matter how hot it gets, he’s not afraid to fight for America.” Trump lies down in the sand, and his micropenis, which has googly eyes and a mouth, slowly becomes erect, before announcing, “I’m Donald J. Trump, and I endorse this message.” The P.S.A. is labelled one of fifty, leaving open the possibility that, in the course of the forty-nine “South Park” episodes still to come, we’ll get forty-nine more.

Is this too much? Probably. Yet there’s an age-old tradition of political vulgarity, of which Trump himself is a practitioner—it’s the crux of his appeal.

by Tyler Foggatt, New Yorker | Read more:
Image:South Park Studios/YouTube

[ed. Classic.]

Let Them Eat Golf Balls

President Trump is using $10 million of our taxes to market his new golf course in Scotland.

The president traveled to Scotland on Friday for the grand opening of an 18-hole golf course in Aberdeen. He’s expected to stay for four days. His appearance will likely generate positive revenue and publicity for the course—money that will flow right back into the pockets of the Trump Organization.

HuffPost has estimated that the trip will likely cost at least $9.7 million dollars due to Air Force One operations, motorcades and helicopters, Secret Service overtime, and more. Trump has framed the international vacation as a “working trip,” and has instead emphasized his plan to meet in Aberdeen with U.K. Prime Minister Kier Starmer. But Aberdeen is not the capital of the United Kingdom, or even the capital of Scotland, making it clear this meeting was just randomly added in to use as an excuse for the golf course.

Trump has grown more and more comfortable completely blurring the lines of his private businesses and his public office. This trip will make his second-term golf tab at least $52 million in just six months, according to HuffPost. His first term was $152 million over four years.

by Malcolm Ferguson, Yahoo News | Read more:
Image: uncredited
[ed. Remember when indignant wingnuts used to turn ten shades of purple whenever Obama went golfing? Must have taken up the sport. See also: this and this. Bonus: Did Trump Cheat at Golf? See the Video For Yourself. (Yahoo News). Of course he did.]

*** Est. cost to taxpayers for golf since returning to office (updated): $68,600,000 - 49 out of 196 days (25.0% of the presidency spent golfing). [ed. ... and counting.]

 (via)

Trump Meets With Powell at Federal Reserve

... leading to one of the most surreal political moments in recent memory

[ed. Trump tries to jack up Federal Reserve Chairman Jerome Powell over supposed cost overuns with fed building rennovations - a clumsy attempt to intimidate over interest rate policy. Normally the theatrical possibilities of something like this would seem about zero. But no! Here we have two of the most powerful people in the world, hard hats precariously balanced on heads, standing in suits in a basement somewhere (presumably the fed's but who knows), arguing over details of a construction contract. Like they say, you can't make this stuff up.]

From Babylon to Wall Street – How Bankers Make You Poor

Michael Hudson has been expanding his historical window, from the ancient history of abolition of debt jubilees, which had prevented the rise of oligarchs, to the increased power over times of creditors, or in lay parlance, bankers. He’s added in the re-establishment of the influence of lenders in medieval times, thanks to the role of the Catholic Church in the Crusades and the accompanying rise of banking to provide war finance. This interview with Jonathan Brown reviews this trajectory, focusing on the way that debt burdens rise over time and amount to destructive rentierism.

Jonathan: You’ve often spoken about your aha [00:01:00] moments when delving into ancient economic history. I just wonder what have been some of your profound or unexpected discoveries about studying ancient civilizations like Sumer or Babylonia?

Michael Hudson: My, entire life, ever since I became an economist in, the 1960s was to realize that debt was the major problem that was going to be growing exponentially and stifling society. And it was clear that debt grew at compound interest faster than the economy was able to grow and pay the debts.

I spent, quite a few, decades warning about the fact that the global south could not pay the Dollarized debts, as indeed it didn’t in the 1970s. There was such a reaction to what I was saying, such a refusal by the economics profession to look at debt as being important, that I decided [00:02:00] to look at the whole history of how different societies had coped with debts.

And I began to write a history of debt, after I left the United Nations in 1979 after warning that there was going to be a, third world, Latin American debt crash in a few years, as indeed there was in 1982. I got all the way back to Greece and Rome, and then into the biblical, and came across the jubilee year. (...)

So I began to write up my ideas, shared them with a friend of mine, Alex Marshak, a professor at Harvard. He introduced me to the head of Harvard’s anthropology and archeology department. I was made a research fellow at the Peabody Museum by Carl Lambert Klowski. I realized that there was this wealth of Babylonian, Sumerian, and near [00:04:00] Eastern, academic records that economists had completely ignored.

And the reasons that our economists ignored it was that the way that society created its economic relationships were completely different from those that they ended up with after Greece and Rome. And so I realized that I can’t simply write this all up myself because I’m an economist, not an Assyriaologist.

So at Harvard we decided to organize a group of scholars who were specialists in Sumerian, Babylonian, Egyptian, Judaic and other Middle Eastern records and we decided to do three volumes.

by Jonathan Brown and Michael Hudson, Naked Capitalism |  Read more:

Image: via
[ed.  See also: The Bull Market for Economists Is Over. It’s an Ominous Sign for the Economy (NYT). Thinkng of commissioning some new t-shirts: Cognitive Dissonance is Killing Me ©]

***

"For decades, earning a Ph.D. in economics has been a nearly foolproof path to a lucrative career. Even as bearers of advanced degrees in history, English or anthropology struggled to find gainful employment, the popularity of economics as an undergraduate major created plenty of tenure-track teaching positions, while government agencies snatched up Ph.D. economists in bulk. Those looking for even larger paychecks could turn to tech companies, Wall Street and consulting firms, which bid up the price of economists as if they were a bespoke cryptocurrency.

Last year, the average base salary for newly hired economics professors at major research universities was more than $150,000, according to the American Economic Association, and their compensation swelled to about $200,000 once bonuses and summer pay were included. As recently as the 2023-24 academic year, the employment rate for Ph.D. economists within a few months of graduation was 100 percent, said John Cawley, the chair of the association’s Committee on the Job Market, citing the group’s surveys. Job satisfaction topped 85 percent.

Those glory days seem to be ending. Universities and nonprofits have scaled back hiring amid declining state budgets and federal funding cuts. At the same time, the Trump administration has laid off government economists and frozen hiring for new ones. (...)

Tech companies also have grown stingier, and their need for high-level economists — once seemingly insatiable — has waned. Other firms have slowed hiring in response to the economic uncertainty introduced by President Trump’s tariffs and the possibility that artificial intelligence will replace their workers, even if those workers have a doctoral degree.

“The advent of A.I. is also impacting the market for high-skilled labor,” said Betsey Stevenson, a labor economist at the University of Michigan, in an email. “So the whole thing is kind of a mess.”

Of course, if it were only some egghead economists scrambling to find work, that might be not be terribly consequential. But the same forces bedeviling economists are crimping employment for other highly trained scientists and social scientists, as well as for many recent college graduates, whose jobless rate has been unusually high for an otherwise strong economy.

The drop in government payrolls and federal funding for universities and nonprofits alone is a major problem, since they support two to three times as many jobs for college graduates as for those without degrees. In some cases, workers with Ph.D.s are displacing others with master’s or bachelor’s degrees.

Then there is the potential impact on the country’s future. Marcia McNutt, a geophysicist who is president of the National Academy of Sciences, said a sharp drop in the number of research jobs in the hard sciences and social sciences would send Ph.D.s abroad. Their flight will deprive the government of the brainpower it needs to perform basic functions and leave U.S. firms less innovative and competitive.

“U.S. industry is incredibly dependent on the training that is done in colleges and universities,” Dr. McNutt said. “When the top people go elsewhere, we’ll be left with the B team in America.”

~ The Bull Market for Economists Is Over. It’s an Ominous Sign for the Economy. Norm Scheiber

Silence on SNAP

Poverty and hunger will rise as a result of the Trump administration’s unprecedented cuts to the US federal “food stamps” program, according to experts. Low-income workers who rely on the aid are braced for dire consequences.

Katie Giede, a single mother and waitress in Conyers, Georgia, is one of the 42 million Americans who use the supplemental nutrition assistance program (Snap). Even with the maximum benefit permitted, she struggles to afford food for her and her child.

She makes $3 an hour plus tips at the fast-food chain Waffle House, where she has worked for 11 years. The company deducts meals from workers’ pay check per shift, regardless of whether they eat one or not.

“Our pay is already so little that we’re struggling with everything,” Giede told the Guardian. “Single mothers like myself are reliant upon the benefits like Snap and Medicaid. So when you go and you cut that as well, now you have mothers out here that are not only worried at night because they already can’t afford housing or a vehicle, but we’re also worried what is our kid is going to eat? Because we no longer have help.”

Giede said she received $450 a month for her and her child. She said working too many hours or receiving too much income was a constant concern, due to eligibility cut-offs.

According to an analysis by the Urban Institute, at the end of 2024, even the maximum Snap benefit would not cover the cost of a modestly priced meal in 99% of all counties in the US.

“I dread that trip to the grocery store every week, because you have to sit down and you really have to budget,” said Giede. “Every time you go, you’re having to make the choice between something that’s healthy or something that’s cheaper, just so you can get enough to last all week.

“There are so many people in this country that rely on these benefits, and with these cuts, half of the people that are surviving right now off of this are going to lose their benefits. That’s not even just people not eating a little bit. They’re already not eating enough, so we’re going to lose lives over this. It’s those of us at the bottom that are really feeling it.”

Waffle House did not respond to multiple requests for comment.

Trump’s “big, beautiful bill” set the stage for significant cuts to Snap by shifting higher administrative costs to each state, expanding work reporting requirements and imposing restrictions on non-citizen eligibility.

Many lower-wage workers have grown more reliant on Snap in recent years. US food prices rose by 23.6% between 2020 and 2024, according to official data. While inflation has since moderated, grocery costs remain high.

As a result of the latest Snap changes, states will be responsible for 75% of administrative costs of handling the program from 2027, up from 50% cost-sharing with the federal government, which is likely to strain state budgets.

From 2028, for the first time states will be forced to pick up some of the multibillion-dollar bill for Snap benefits. The state of New York, for example, faces a budget impact of about $1.2bn, according to the Food Research and Action Center (Frac), a non-profit advocacy group.

While such shifting costs have raised fears that states will cut back Snap support, expanded work requirements have sparked concern that few people will be eligible. Analysis by the Urban Institute found about 22.3 million US families are set to lose some or all of their Snap benefits.

“This is a very targeted, well-thought-out plan of dismantling the Snap program that federal policy makers won’t take responsibility for, because it is the states, it is the governors who will have to cut resources for Snap, who will have to cut the program in order to say we can’t operate this because of what’s happening at the federal level,” said Gina Plata-Nino, Snap deputy director at the Frac.

“Snap is a very important ecosystem at the local level, at the state level and the federal level, because billions of dollars go into states, and this federal money supports local economies,” she added. “All of these proposals threaten this very delicate balance.”

The White House deferred comment to the office of management and budget, which did not respond to multiple requests for comment.

States across the US are braced for stark consequences. “We’re going to have worse hunger and ultimately, worse poverty,” said Seth DiStefano, policy outreach director at the West Virginia Center on Budget and Policy. “There are entire regions of West Virginia where there aren’t 20 hours a week [expanded Snap work requirement] of anything to apply for. What do you tell those families?

“We’re talking families with kids now that are going to be subjected to these harsh work reporting requirements. We’re talking folks in their 60s, literally in communities where there are no jobs, none, and ripping away the one outlet to their basic needs that’s available to them.”

Among the employers with the most workers reliant on Snap is Walmart, the largest private employer in the US, as much of its workforce receives only part-time hours.

Christina Gahagan, 66, has worked at Walmart for a decade in western New York at several stores. She is currently based at a store in Geneseo, New York.

“I would say at least 50% of the people in my store rely on food stamps to make ends meet for their families,” said Gahagan. “They’re always trying to figure out where the best deals are, coupon clipping at lunch and reading circulars to see who’s got the best deal on whatever, just to make their money stretch.” (...)

“Walmart is the largest employer in the US. We rival Amazon almost dollar for dollar in what we do. You would think a company like that could shell out a little bit more money per hour for associates in the store across the board, so that there aren’t people who are having to depend so heavily on public assistance.”

Walmart did not respond to multiple requests for comment.

by Michael Sainato, The Guardian | Read more:
Image:Richard Levine/Alamy
[ed. No one wants to comment on a new bureaucracy to process and administer oversight requirements? Jobs! Remember who did this the next time you vote.]

via:

[ed. absolutely..]

Brigitte Yoshiko Pruchnow, Greek Salad, 2024
via:

Design Your Own Rug!

For my wedding anniversary, I designed and had hand-woven in Afghanistan a rug for my microbiologist wife. The rug mixes traditional Afghanistan designs with some scientific elements including Bunsen burners, test tubes, bacterial petri dishes and other elements.

I started with several AI designs, such as that shown below, to give the weavers an idea of what I was looking for. Some of the AI elements were muddled and very complex and so we developed a blueprint over a few iterations. The blueprint was very accurate to the actual rug.

I am very pleased with the final product. The wool is of high quality, deep and luxurious, and the design is exactly what I intended. My wife loves the rug and will hang it at her office. The price was very reasonable, under $1000. I also like that I employed weavers in a small village in Northern Afghanistan. The whole process took about 6 months.

You can develop your own custom rug from Afghanu Rugs. Tell them Alex sent you. Of course, they also have many beautiful traditional designs. You can even order my design should you so desire!

by Alex Tabarrok, Marginal Revolution | Read more:
Images: the author

Breakneck: China’s Quest to Engineer the Future

I didn’t write a letter last year. Rather, I wrote seven, all of which is new material.

They make up my book BREAKNECK: China’s Quest to Engineer the Future. It’s driven by a few simple ideas. That Americans and Chinese are fundamentally alike: restless, eager for shortcuts, ultimately driving most of the world’s big changes. That their rivalry should not be reasoned through with worn-out terms from the past century like socialist, democratic, or neoliberal. And that both countries are tangles of imperfection, regularly delivering — in the name of competition — self-beatings that go beyond the wildest dreams of the other.

The simplest idea I present is that China is an engineering state, which brings a sledgehammer to problems both physical and social, in contrast with America’s lawyerly society, which brings a gavel to block almost everything, good and bad.

Breakneck begins with a bike ride I took from Guiyang to Chongqing in 2021. China’s fourth-poorest province, I was delighted to find, has much better infrastructure than California or New York, both wealthier by orders of magnitude. Five days of grueling climbs on stunning green mountains gave me glimpses of what socialism with Chinese characteristics really looks like. But there is more to the engineering state than tall bridges. The heart of the book concerns how badly Beijing goes off track when it engages in social engineering. My handy formulation of the Communist Party is that it is a Leninist Technocracy with Grand Opera Characteristics — practical until it collapses into the preposterous.

The idea of the lawyerly society became obvious when I returned to the U.S. in 2023. The Paul Tsai China Center (as I say in my acknowledgments) was the best possible place to write this book, not only because it’s so supportive, but also because it set me inside the Yale Law School. Elite law schools, now and in the past, fashion the easiest path for the ambitions to step into the top ranks of the American government. The dominance of lawyers in the American elite has helped transmute the United States into a litigious vetocracy. I believe that America cannot remain a great power if it is so committed to a system that works well mostly for the wealthy and well-connected.

The engineering state versus the lawyerly society is not a grand theory to explain absolutely everything about the U.S. and China. Rather, the book is rooted in my own experiences of living in China from 2017 to 2023. I offer this framework to make sense of the recent past and think about what might come next.

It helps to explain a number of things. For example, the trade war and the tech showdown. The U.S. has relied on legalisms — levying tariffs and designing an ever more exquisite sanctions regime — while China has focused on creating the future by physically building better cars, more beautiful cities, and bigger power plants. Though China has constructed roads and bridges abroad, it struggles to inspire global cultural appeal, because engineers aren’t smooth talkers and tend to censor whatever they can’t understand. The Chinese state is sometimes too rational, proceeding down a path that feels perfectly logical, until the country’s largest city is suddenly in a state of lockdown for months.

Breakneck will be published on August 26. I hope you’ll order this book. You can also send me an email if you would like a review copy for your publication or Substack, or to book me for speaking.

***

It’s a bit boring to write only a book announcement. This is also a space for me to reflect on the bookwriting process.

The hard part of bookwriting is the beginning, the middle, and the end. Each stage demands unrelated skills. The opening phases involve engaging an agent, beating ideas into the shape of a proposal (which typically stretch over 50 pages), and approaching a publisher. The long middle is the writing. The end is the mishmash of tasks related to revision, production, and promotion. Fortunately I had a superb agent and a faithful editor to navigate the first and third stages. Overall the process was more fun than I expected, such that I now actively encourage friends to pursue their own book ideas.

Writing is necessarily a solitary task. My usual process is to putter around until late evening, until I finally cannot bear to avoid the page any longer, at which point I spend a lot of time picking out appropriate music, and finally get to the task. I knew that could no longer be a sane approach for a lengthier writing project (not that it ever was). Every day I repeated my mantra to be a cool, calm, collected Canadian, through which I achieved a modest degree of discipline. I met my deadline.

I became a better writer over the course of the book. Breakneck, as I said, is seven annual letters. I thought I understood this format, but I still saw myself improving, such that the final chapter was much easier to write than the first. I felt my prose loosening and my confidence rising as I moved from chapter to chapter. Bookwriting is a bit like climbing a mountain: best not to look up too much at the beginning and feel daunted by the task ahead. When I had completed two-thirds of the book, I started feeling elated about how much I’ve written, which propelled me towards the end.

Writing is thinking. As I worked on my final chapter, I found myself reflecting on my Yunnan heritage. Yunnan is, in my estimation, China’s freest province: far away amid southwestern mountains, it has mostly escaped sustained attention from the imperial center, which would be attracted to greater wealth or restive minority issues. My parents both have deep Yunnan roots. They would have been in China’s middle class, only the concept did not really exist when they emigrated to Canada when I was seven. I’m glad to have had an upbringing in this economic backwater, which is undeveloped in part because it’s inflected by a bit of the suspicion of the state that is common to mountain peoples everywhere. Growing up in the periphery endowed me with greater skepticism of the state glories that Beijing chooses to celebrate and greater reluctance to participate in the competitive culture common in Shanghai or Shenzhen.

I wrote this book partly to sort out my own thoughts about China. It really was staggering to write about how many miles of roadways, how many new nuclear power plants, how much steel China has produced over the past four decades. China is a good operating model of abundance. I state clearly in the book that America doesn’t have to become China to build infrastructure; it would be sufficient to reach the construction cost levels of France, Japan, or Spain. Still, the U.S. should still study some aspects of China’s method: how do they build it? What are the tradeoffs? How do we learn? China has gotten a lot of things right with mass transit, plentiful housing, and functional cities.

The problem is that China’s leadership just can’t stop at physical engineering. Sooner or later, they treat the population as if it were another building material, to be moulded or torn apart as the circumstances demand. That’s why America shouldn’t look to China as the model. My favorite chapter concerned the one-child policy. I had been completely unprepared to study the brutality of its enforcement, which was only possible through mass sterilizations and forced abortions. At its peak in the 1980s, the one-child policy morphed into a campaign of rural terror meted out against female bodies, namely the mother and the cruelly discarded daughter.

Nearly all the letters are focused on China. The final one is about the United States. I concluded my book by writing about what my parents gained and lost with their emigration. They lost the chance to build wealth as part of China’s luckiest generation: urban residents born after 1960 who were able to acquire property or build businesses after the 2000s. But they would not trade that for their gain of living in the suburbs of Philly, which I find boring, but their friends find enviable. I also reflected on America’s own legacy as an engineering state, focused on two engineers: Robert Moses and Hyman Rickover. Too many parts of America feel like the well-preserved ruins of a once-great civilization. Americans should take a clearer look at the industrial achievements that are usually ignored and frequently scorned. (...)

I cooked a lot of fish as I wrote, in the Cantonese style: steaming a whole bronzino or a filet of sea trout for ten minutes, then drizzled with ginger, spring onion, soy sauce, and sizzling olive oil. My wife and I also planned a few writing retreats, in which we would park ourselves in new places to focus on food, exercise, and writing. After six years of intensively eating Chinese cuisines, I was also pleased to move into new culinary worlds. (...)

When I last visited Shanghai, at the end of 2024, I was surprised to feel that the average person might be eating worse than before. The trend of consumption downgrading has been real. Smart restaurants are no longer difficult to book. Sichuan and Hunan restaurants are taking over. A lot of the restaurant foods are prepared in centralized commissaries. Many more places focus more on deliveries than the sit-down experience. And there seems to be a trend of chain restaurants from third-tier cities moving to first-tier cities, offering slightly worse food at much cheaper prices.

The worst part is the influencer culture. China’s influencer culture is much more intense than America’s. It’s easy to see, in public spaces, how many people are glued to their phones. Anywhere charming, whether a café or a mountaintop, is full of people intently taking photos. It’s common to see Chinese couples or groups of friends barely interacting with each other over a meal, leaning over their phones. I remember having coffee once at the Ritz-Carlton in Shanghai, where a group of girls sat near me photographing each other over cakes for over an hour. Influencer culture has pushed restaurants to make dishes better photographed than tasted.

It doesn’t mean that China will fall behind America in food. No way. China retains a commanding lead, and it has so much vitality in smaller cities and the countryside. But I wonder whether China will maintain its culinary peaks, or if they will be corroded by consumer-driven homogenization and the priority of convenience over tastiness. On present trendlines, America is learning to get better, while China is slightly worse.

by Dan Wang |  Read more:

Image: Breakneck

[ed. I've been a fan of Dan's annual China summaries since discovering them back in 2021 (see here, here and here). When 2025 rolled around and none appeared I wrote and asked if he was still planning something. That's when he told me about this book. Definitely plan to get it when it's released.]