Understandary

•

Canal • 2 podcasts

Comentários de notícias

Comentários de notícias

Semanal
Sociedade e cultura

Sociedade e cultura

Quinzenal

23/12/2025

Data Center Politics

This week we talk about energy consumption, pollution, and bipartisan issues. We also discuss local politics, data center costs, and the Magnificent 7 tech companies. Recommended Book: Against the Machine by Paul Kingsnorth Transcript In 2024, the International Energy Agency estimated that data centers consumed about 1.5% of all electricity generated, globally, that year. It went on to project that energy consumption by data centers could double by 2030, though other estimates are higher, due to the ballooning of investment in AI-focused data centers by some of the world’s largest tech companies. There are all sorts of data centers that serve all kinds of purposes, and they’ve been around since the mid-20th century, since the development of general purposes digital computers, like the 1945 Electronic Numerical Integrator and Computer, or ENIAC, which was programmable and reprogrammable, and used to study, among other things, the feasibility of thermonuclear weapons. ENIAC was built on the campus of the University of Pennsylvania and cost just shy of $500,000, which in today’s money would be around $7 million. It was able to do calculators about a thousand times faster than other, electro-mechanical calculators that were available at the time, and was thus considered to be a pretty big deal, making some types of calculation that were previously not feasible, not only feasible, but casually accomplishable. This general model of building big-old computers at a center location was the way of things, on a practical level, until the dawn of personal computers in the 1980s. The mainframe-terminal setup that dominated until then necessitated that the huge, cumbersome computing hardware was all located in a big room somewhere, and then the terminal devices were points of access that allowed people to tap into those centralized resources. Microcomputers of the sort of a person might have in their home changed that dynamic, but the dawn of the internet reintroduced something similar, allowing folks to have a computer at home or at their desk, which has its own resources, but to then tap into other microcomputers, and to still other larger, more powerful computers across internet connections. Going on the web and visiting a website is basically just that: connecting to another computer somewhere, that distant device storing the website data on its hard drive and sending the results to your probably less-powerful device, at home or work. In the late-90s and early 2000s, this dynamic evolved still further, those far-off machines doing more and more heavy-lifting to create more and more sophisticated online experiences. This manifested as websites that were malleable and editable by the end-user—part of the so-called Web 2.0 experience, which allowed for comments and chat rooms and the uploading of images to those sites, based at those far off machines—and then as streaming video and music, and proto-versions of social networks became a thing, these channels connecting personal devices to more powerful, far-off devices needed more bandwidth, because more and more work was being done by those powerful, centrally located computers, so that the results could be distributed via the internet to all those personal computers and, increasingly, other devices like phones and tablets. Modern data centers do a lot of the same work as those earlier iterations, though increasingly they do a whole lot more heavy-lifting labor, as well. They’ve got hardware capable of, for instance, playing the most high-end video games at the highest settings, and then sending, frame by frame, the output of said video games to a weaker device, someone’s phone or comparably low-end computer, at home, allowing the user of those weaker devices to play those games, their keyboard or controller inputs sent to the data center fast enough that they can control what’s happening and see the result on their own screen in less than the blink of an eye. This is also what allows folks to store backups on cloud servers, big hard drives located in such facilities, and it’s what allows the current AI boom to function—all the expensive computers and their high-end chips located at enormous data centers with sophisticated cooling systems and high-throughput cables that allow folks around the world to tap into their AI models, interact with them, have them do heavy-lifting for them, and then those computers at these data centers send all that information back out into the world, to their devices, even if those devices are underpowered and could never do that same kind of work on their own. What I’d like to talk about today are data centers, the enormous boom in their construction, and how these things are becoming a surprise hot button political issue pretty much everywhere. — As of early 2024, the US was host to nearly 5,400 data centers sprawled across the country. That’s more than any other nation, and that number is growing quickly as those aforementioned enormous tech companies, including the Magnificent 7 tech companies, Nvidia, Apple, Alphabet, Microsoft, Amazon, Meta, and Tesla, which have a combined market cap of about $21.7 trillion as of mid-December 2025, which is about two-thirds of the US’s total GDP for the year, and which is more than the European Union’s total GDP, which weighs in at around $19.4 trillion, as of October 2025—as they splurge on more and more of them. These aren’t the only companies building data centers at breakneck speed—there are quite a few competitors in China doing the same, for instance—but they’re putting up the lion’s share of resources for this sort of infrastructure right now, in part because they anticipate a whole lot of near-future demand for AI services, and those services require just a silly amount of processing power, which itself requires a silly amount of monetary investment and electricity, but also because, first, there aren’t a lot of moats, meaning protective, defensive assets in this industry, as is evidenced by their continual leapfrogging of each other, and the notion that a lot of what they’re doing, today, will probably become commodity services in not too long, rather than high-end services people and businesses will be inclined to pay big money for, and second, because there’s a suspicion, held by many in this industry, that there’s an AI shake-out coming, a bubble pop or bare-minimum a release of air from that bubble, which will probably kill off a huge chunk of the industry, leaving just the largest, too-big-to-fail players still intact, who can then gobble up the rest of the dying industry at a discount. Those who have the infrastructure, who have invested the huge sums of money to build these data centers, basically, will be in a prime position to survive that extinction-level event, in other words. So they’re all scrambling to erect these things as quickly as possible, lest they be left behind. That construction, though, is easier said than done. The highest-end chips account for around 70-80% of a modern data center’s cost, as these GPUs, graphical processing units that are optimized for AI purposes, like Nvidia’s Blackwell chips, can cost tens of thousands of dollars apiece, and millions of dollars per rack. There are a lot of racks of such chips in these data centers, and the total cost of a large-scale AI-optimized data center is often somewhere between $35 and $60 billion. A recent estimate by McKinsey suggests that by 2030, data center investment will need to be around $6.7 trillion a year just to keep up the pace and meet demand for compute power. That’s demand from these tech companies, I should say—there’s a big debate about where there’s sufficient demand from consumers of AI products, and whether these tech companies are trying to create such demand from whole cloth, to justify heightened valuations, and thus to continue goosing their market caps, which in turn enriches those at the top of these companies. That said, it’s a fair bet that for at least a few more years this influx in investment will continue, and that means pumping out more of these data centers. But building these sorts of facilities isn’t just expensive, it’s also regulatorily complex. There are smaller facilities, akin to ENIAC’s campus location, back in the day, but a lot of them—because of the economies of scale inherent in building a lot of this stuff all at once, all in the same place—are enormous, a single data center facility covering thousands of acres and consuming a whole lot of power to keep all of those computers with their high-end chips running 24/7. Previous data centers from the pre-AI era tended to consume in the neighborhood of 30MW of energy, but the baseline now is closer to 200MW. The largest contemporary data centers consume 1GW of electricity, which is about the size of a small city’s power grid—that’s a city of maybe 500,000-750,000 people, though of course climate, industry, and other variables determine the exact energy requirements of a city—and they’re expected to just get larger and more resource-intensive from here. This has resulted in panic and pullbacks in some areas. In Dublin, for instance, the government has stopped issuing new grid connections for data centers until 2028, as it’s estimated that data centers will account for 28% of Ireland’s power use by 2031, already. Some of these big tech companies have read the writing on the wall, and are either making deals to reactivate aging power plants—nuclear, gas, coal, whatever they can get—or are saying they’ll build new ones to offset the impact on the local power grid. And that impact can be significant. In addition to the health and pollution issues caused by some of the sites—in Memphis, for instance, where Elon Musk’s company, xAI, built a huge data center to help power his AI chatbot, Grok, the company is operating 35 unpermitted gas turbi

23/12/2025

•

17min
16/12/2025

Chip Exports

This week we talk about NVIDIA, AI companies, and the US economy. We also discuss the US-China chip-gap, mixed-use technologies, and export bans. Recommended Book: Enshittification by Cory Doctorow Transcript I’ve spoken about this a few times in recent months, but it’s worth rehashing real quick because this collection of stories and entities are so central to what’s happening across a lot of the global economy, and is also fundamental, in a very load-bearing way, to the US economy right now. As of November of 2025, around the same time that Nvidia, the maker of the world’s best AI-optimized chips at the moment became the world’s first company to achieve a $5 trillion market cap, the top seven highest-valued tech companies, including Nvidia, accounted for about 32% of the total value of the US stock market. That’s an absolutely astonishing figure, as while Nvidia, Apple, Microsoft, Alphabet, Amazon, Broadcom, and Meta all have a fairly diverse footprint even beyond their AI efforts, a lot of that value for all of them is predicated on expected future income; which is to say, their market caps, their value according to that measure, is determined not by their current assets and revenue, but by what investors think or hope they’ll pull in and be worth in the future. That’s important to note because historically the sorts of companies that have market caps that are many multiples of their current, more concrete values are startups; companies in their hatchling phase that have a good idea and some kind of big potential, a big moat around what they’re offering or a blue ocean sub-industry with little competition in which they can flourish, and investment is thus expected to help them grow fast. These top seven tech companies, in contrast, are all very mature, have been around for a while and have a lot of infrastructure, employees, expenses, and all the other things we typically associated with mature businesses, not flashy startups with their best days hopefully ahead of them. Some analysts have posited that part of why these companies are pushing the AI thing so hard, and in particular pushing the idea that they’re headed toward some kind of generally useful AI, or AGI, or superhuman AI that can do everyone’s jobs better and cheaper than humans can do them, is that in doing so, they’re imagining a world in which they, and they alone, because of the costs associated with building the data centers required to train and run the best-quality AI right now, are capable of producing basically an economy’s-worth of AI systems and bots and machines operated by those AI systems. In other words, they’re creating, from whole cloth, an imagined scenario in which they’re not just worthy of startup-like valuations, worthy of market caps that are tens or hundreds of times their actual concrete value, because of those possible futures they’re imagining in public, but they’re the only companies worthy of those valuation multiples; the only companies that matter anymore. It’s likely that even if this is the case, that the folks in charge of these companies, and the investors who have money in them who are likely to profit when the companies grow and grow, actually do believe what they’re telling everyone about the possibilities inherent in building these sorts of systems. But there also seems to be a purely economic motive for exaggerating a lot and clearing out as much of the competition as possible as they grow bigger and bigger. Because maybe they’ll actually make what they’re saying they can make as a result of all that investment, that exuberance, but maybe, failing that, they’ll just be the last companies standing after the bubble bursts and an economic wildfire clears out all the smaller companies that couldn’t get the political relationships and sustaining cash they needed to survive the clear-out, if and when reality strikes and everyone realizes that sci-fi outcome isn’t gonna happen, or isn’t gonna happen any time soon. What I’d like to talk about today is a recent decision by the US government to allow Nvidia to sell some of its high-powered chips to China, and why that decision is being near-universally derided by those in the know. — In early December 2025, after a lot of back-and-forthing on the matter, President Trump announced that the US government will allow Nvidia, which is a US-based company, to export its H200 processors to China. He also said that the US government will collect a 25% fee on these sales. The H200 is Nvidia’s second-best chip for AI purposes, and it’s about six-times as powerful as the H20, which is currently the most advanced Nvidia chip that’s been cleared for sale to China. The Blackwell chip that is currently Nvidia’s most powerful AI offering is about 1.5-times faster than the H200 for training purposes, and five-times faster for AI inferencing, which is what they’re used for after a model is trained, and then it’s used for predictions, decisions, and so on. The logic of keeping the highest-end chips from would-be competitors, especially military competitors like China, isn’t new—this is something the US and other governments have pretty much always done, and historically even higher-end gaming systems like Playstation consoles have been banned for export in some cases because the chips they contained could be repurposed for military things, like plucking them out and using them to guide missiles—Sony was initially unable to sell the Playstation 2 outside of Japan because it needed special permits to sell something so militarily capable outside the country, and it remained unsellable in countries like Iraq, Iran, and North Korea throughout its production period. The concern with these Nvidia chips is that if China has access to the most powerful AI processors, it might be able to close the estimated 2-year gap between US companies and Chinese companies when it comes to the sophistication of their AI models and the power of their relevant chips. Beyond being potentially useful for productivity and other economic purposes, this hardware and software is broadly expected to shape the next generation of military hardware, and is already in use for all sorts of wartime and defense purposes, including sophisticated drones used by both sides in Ukraine. If the US loses this advantage, the thinking goes, China might step up its aggression in the South China Sea, potentially even moving up plans to invade Taiwan. Thus, one approach, which has been in place since the Biden administration, has been to do everything possible to keep the best chips out of Chinese hands, because that would ostensibly slow them down, make them less capable of just splurging on the best hardware, which they could then use to further develop their local AI capabilities. This approach, however, also incentivized the Chinese government to double-down on their own homegrown chip industry. Which again is still generally thought to be about 2-years behind the US industry, but it does seem to be closing the gap rapidly, mostly by copying designs and approaches used by companies around the world. An alternative theory, the one that seems to be at least partly responsible for Trump’s about-face on this, is that if the US allows the sale of sufficiently powerful chips to China, the Chinese tech industry will become reliant on goods provided by US companies, and thus its own homegrown AI sector will shrivel and never fully close that gap. If necessary the US can then truncate or shut down those shipments, crippling the Chinese tech industry at a vital moment, and that would give the US the upper-hand in many future negotiations and scenarios. Most analysts in this space no longer think this is a smart approach, because the Chinese government is wise to this tactic, using it itself all the time. And even in spaces where they have plenty of incoming resources from elsewhere, they still try to shore-up their own homegrown versions of the same, copying those international inputs rather than relying on them, so that someday they won’t need them anymore. The same is generally thought to be true, here. Ever since the first Trump administration, when the US government started its trade war with China, the Chinese government has not been keen on ever relying on external governments and economies again, and it looks a lot more likely, based on what the Chinese government has said, and based on investments across the Chinese market on Chinese AI and chip companies following this announcement, that they’ll basically just scoop up as many Nvidia chips as they can, while they can, and primarily for the purpose of reverse-engineering those chips, speeding up their gap-closing with US companies, and then, as soon as possible, severing that tie, competing with Nvidia rather than relying on it. This is an especially pressing matter right now, then, because the US economy, and basically all of its growth, is so completely reliant on AI tech and the chips that are allowing that tech to move forward. If this plan by the US government doesn’t pan out and ends up being a short-term gain situation, a little bit of money earned from that 25% cut the government takes, and Ndvidia temporarily enriching itself further through Chinese sales, but in exchange both entities give up their advantage, long term, to Chinese AI companies and the Chinese government, that could be bad not just for AI companies around the world, which could be rapidly outcompeted by Chinese alternatives, but also all economies exposed to the US economy, which could be in for a long term correction, slump, or full-on depression. Show Notes https://www.nytimes.com/2025/12/09/us/politics/trump-nvidia-ai-chips-china.html https://arstechnica.com/tech-policy/2025/12/us-taking-25-cut-of-nvidia-chip-sales-makes-no-sense-experts-say/ https://www.pcmag.com/news/20-years-later-how-concerns-about-weaponized-consoles-a

16/12/2025

•

14min
02/12/2025

Climate Risk

This week we talk about floods, wildfires, and reinsurance companies. We also discuss the COP meetings, government capture, and air pollution. Recommended Book: If Anyone Builds It, Everyone Dies by Eliezer Yudkowsky and Nate Soares Transcript The urban area that contains India’s capital city, New Delhi, called the National Capital Territory of Delhi, has a population of around 34.7 million people. That makes it the most populous city in the country, and one of the most populous cities in the world. Despite the many leaps India has made over the past few decades, in terms of economic growth and overall quality of life for residents, New Delhi continues to have absolutely abysmal air quality—experts at India’s top research hospital have called New Delhi’s air “severe and life-threatening,” and the level of toxic pollutants in the air, from cars and factories and from the crop-waste burning conducted by nearby farmers, can reach 20-times the recommended level for safe breathing. In mid-November 2025, the problem became so bad that the government told half its workers to work from home, because of the dangers represented by the air, and in the hope that doing so would remove some of the cars on the road and, thus, some of the pollution being generated in the area. Trucks spraying mist, using what are called anti-smog guns, along busy roads and pedestrian centers help—the mist keeping some of the pollution from cars from billowing into the air and becoming part of the regional problem, rather than an ultra-localized one, and pushing the pollutants that would otherwise get into people’s lungs down to the ground—though the use of these mist-sprayers has been controversial, as there are accusations that they’re primarily deployed near air-quality monitoring stations, and that those in charge put them there to make it seem like the overall air-quality is lower than it is, manipulating the stats so that their failure to improve practical air-quality isn’t as evident. And in other regional news, just southeast across the Bay of Bengal, the Indonesian government, as of the day I’m recording this, is searching for the hundreds of people who are still missing following a period of unusually heavy rains. These rains have sparked floods and triggered mudslides that have blocked roads, damaged bridges, and forced the evacuation of entire villages. More than 300,000 people have been evacuated as of last weekend, and more rain is forecast for the coming days. The death toll of this round of heavy rainfall—the heaviest in the region in years—has already surpassed 440 people in Indonesia, with another 160 and 90 in Thailand and Vietnam, respectively, being reported by those countries’ governments, from the same weather system. In Thailand, more than two million people were displaced by flooding, and the government had to deploy military assets, including helicopters launched from an aircraft carrier, to help rescue people from the roofs of buildings across nine provinces. In neighboring Malaysia, tens of thousands of people were forced into shelters as the same storm system barreled through, and Sri Lanka was hit with a cyclone that left at least 193 dead and more than 200 missing, marking one of the country’s worst weather disasters in recent years. What I’d like to talk about today is the climatic moment we’re at, as weather patterns change and in many cases, amplify, and how these sorts of extreme disasters are also causing untold, less reported upon but perhaps even more vital, for future policy shifts, at least, economic impacts. — The UN Conference of the Parties, or COP meetings, are high-level climate change conferences that have typically been attended by representatives from most governments each year, and where these representatives angle for various climate-related rules and policies, while also bragging about individual nations’ climate-related accomplishments. In recent years, such policies have been less ambitious than in previous ones, in part because the initial surge of interest in preventing a 1.5 degrees C increase in average global temperatures is almost certainly no longer an option; climate models were somewhat accurate, but as with many things climate-related, seem to have actually been a little too optimistic—things got worse faster than anticipated, and now the general consensus is that we’ll continue to shoot past 1.5 degrees C over the baseline level semi-regularly, and within a few years or a decade, that’ll become our new normal. The ambition of the 2015 Paris Agreement is thus no longer an option. We don’t yet have a new, generally acceptable—by all those governments and their respective interests—rallying cry, and one of the world’s biggest emitters, the United States, is more or less absent at new climate-related meetings, except to periodically show up and lobby for lower renewables goals and an increase in subsidies for and policies that favor the fossil fuel industry. The increase in both number and potency of climate-influenced natural disasters is partly the result of this failure to act, and act forcefully and rapidly enough, by governments and by all the emitting industries they’re meant to regulate. The cost of such disasters is skyrocketing—there are expected to be around $145 billion in insured losses, alone, in 2025, which is 6% higher than in 2024—and their human impact is booming as well, including deaths and injuries, but also the number of people being displaced, in some cases permanently, by these disasters. But none of that seems to move the needle much in some areas, in the face of entrenched interests, like the aforementioned fossil fuel industry, and the seeming inability of politicians in some nations to think and act beyond the needs of their next election cycle. That said, progress is still being made on many of these issues; it’s just slower than it needs to be to reach previously set goals, like that now-defunct 1.5 degrees C ceiling. Most nations, beyond petro-states like Russia and those with fossil fuel industry-captured governments like the current US administration, have been deploying renewables, especially solar panels, at extraordinary rates. This is primarily the result of China’s breakneck deployment of solar, which has offset a lot of energy growth that would have otherwise come from dirty sources like coal in the country, and which has led to a booming overproduction of panels that’s allowed them to sell said panels cheap, overseas. Consequently, many nations, like Pakistan and a growing number of countries across Sub-Saharan African, have been buying as many cheap panels as they can afford and bypassing otherwise dirty and unreliable energy grids, creating arrays of microgrids, instead. Despite those notable absences, then, solar energy infrastructure installations have been increasing at staggering rates, and the first half of 2025 has seen the highest rate of capacity additions, yet—though China is still installing twice as much solar as the rest of the world, combined, at this point. Which is still valuable, as they still have a lot of dirty energy generation to offset as their energy needs increase, but more widely disseminated growth is generally seen to be better in the long-term—so the expansion into other parts of the world is arguably the bigger win, here. The economics of renewables may, at some point, convince even the skeptics and those who are politically opposed to the concept of renewables, rather than practically opposed to them, that it’s time to change teams. Already, conservative parts of the US, like Texas, are becoming renewables boom-towns, quietly deploying wind and solar because they’re often the best, cheapest, most resilient options, even as their politicians rail against them in public and vote for more fossil fuel subsidies. And it may be economics that eventually serve as the next nudge, or forceful shove on this movement toward renewables, as we’re reaching a point at which real estate and the global construction industry, not to mention the larger financial system that underpins them and pretty much all other large-scale economic activities, are being not just impacted, but rattled at their roots, by climate change. In early November 2025, real estate listing company Zillow, the biggest such company in the US, stopped showing extreme weather risks for more than a million home sale listings on its site. It started showing these risk ratings in 2024, using data from a risk-modeling company called First Street, and the idea was to give potential buyers a sense of how at-risk a property they were considering buying might be when it comes to wildfires, floods, poor air quality, and other climate and pollution-related issues. Real estate agents hated these ratings, though, in part because there was no way to protest and change them, but also because, well, they might have an expensive coastal property listed that now showed potential buyers it was flood prone, if not today, in a couple of years. It might also show a beautiful mountain property that’s uninsurable because of the risk of wildfire damage. A good heuristic for understanding the impact of global climate change is not to think in terms of warming, though that’s often part of it, but rather thinking in terms of more radical temperature and weather swings. That means areas that were previously at little or no risk of flooding might suddenly be very at risk of absolutely devastating floods. And the same is true of storms, wildfires, and heat so intense people die just from being outside for an hour, and in which components of one’s house might fry or melt. This move by Zillow, the appearance and removal of these risk scores, happened at the same time global insurers are warning that they may have to pull out of more areas, because it’s simply no longer possible for them to do business in pla

02/12/2025

•

16min
04/11/2025

Supersonic Flight

This week we talk about Mach 1, the Bell X-1, and the Concorde. We also discuss the X-59, the Tu-144, and Boom Supersonic. Recommended Book: Red Team Blues by Cory Doctorow Transcript The term “supersonic,” when applied to speed, refers to something moving faster than the speed of sound—a speed that is shorthanded as Mach 1. The precise Mach 1 speed of sound will be different depending on the nature of the medium through which an object is traveling. So if you’re moving at sea level versus up high in the air, in the stratosphere, the speed of sound will be different. Likewise if you’re moving through moist air versus dry air, or moving through water versus moving through syrup, different speed of sound, different Mach 1. In general, though, to give a basic sense of how fast we’re talking here, if an object is moving at sea level through dry air at a temperature of 20 degrees celsius, which is 68 degrees fahrenheit, Mach 1 is about 768 miles per hour, which is about 1,126 feet per second, and 343.2 meters per second. It’s fast! It’s very fast. Again, this is the speed at which sound moves. So if you surpass the speed of sound, if you go supersonic, you will arrive faster than the sound you make while moving. Back in 1947, an experimental American plane called the Bell X-1 broke the sound barrier, surpassed Mach 1, reaching a speed of almost 1,000 miles per hour using a 6,000 pound thrust rocket propulsion system. A later version of the same rocket-powered plane, the Bell X-1A, which was basically the same vehicle, it just had more fuel capacity, allowing the rocket to burn longer, achieved 1,600 miles per hour in 1956. Prior to that, in 1943, British began working on a secret experimental aircraft called the Miles M.52, intending to build a plane capable of traveling 1,000 mph. Interestingly, this project was apparently the result of the British wanting to keep up with a supposed already existing German aircraft capable of achieving that speed, though it’s now believed the intelligence that led the British to believe the Germans had a supersonic-capable plane was the result of a mistranslation—the Germans hit 1,000 km per hour, which is about 621 mph, and still subsonic. Though apparently a success in terms of research and innovation, the Miles M.52 project was cancelled in 1946, due partly to budgetary concerns, and partly because the new government didn’t believe supersonic aircraft were practical, or maybe even feasible. After the existence of this project was revealed to the public, however, criticism for the cancellation mounted, and the design was translated into new, unmanned scale-model experimental versions of the plane which achieved controlled Mach 1.38 supersonic speeds, and both the design and research from this program was shared with the American company, Bell, and all that knowledge informed the development of the aforementioned Bell X-1 supersonic plane. Again, that successful Bell mission was flown in 1947, and in 1961, a Douglas jetliner, a commercial jet, broke the sound barrier during a controlled test dive, and that fed the development of an intended supersonic airliner in the US, though similar research being conducted elsewhere would bear more direct and immediate fruit. In the Soviet Union, a supersonic jetliner called the Tupolev Tu-144 entered service in 1968, and a jetliner co-developed by the British and French, the Concorde, began construction in 1965, and tallied its first flight in March of 1969. The Tu-144 was thus the world’s first commercial supersonic airliner, by a few months, and it also became the first commercial transport to exceed Mach 2, twice the speed of sound, in 1970. The Tu-144 was plagued by reliability issues from the get-go, however, and while performing maneuvers at an air show in Paris in 1973, it disintegrated in midair, which—combined with its high operating costs reduced its long-term market viability, especially internationally. By the mid-1970s, it was primarily operating within the Soviet Union, and after a new variant of the jet crashed in 1978, the Tu-144 program was cancelled in 1983. Existing models continued to be use for niche purposes, like training space program pilots, and for a supersonic research program undertaken by NASA in the late-1990s, but the final Tu-144 flight was in mid-1999, and all surviving aircraft are now on display or in storage. The Concorde has a similar history. Original forecasts for the supersonic airliner market were optimistic, and while the craft seemed to be generally more reliable and less issue-prone than the Tu-144, and it enjoyed a period of fanfare and promotion, as a sort of luxury experience for folks crossing the Atlantic in particular, cutting travel times in half, a major crash in mid-2000, which killed all 109 occupants and four people on the ground, led to the suspension of service until late-2001, and all remaining Concorde aircraft were retired in 2003—about 20 of them are on display throughout North American and Europe, as of the mid-2020s. The costs associated with operating Concorde aircraft, as with the Tu-144, were also quite high, and those costs and other complications led to the cancellation of a would-be supersonic jetliner competitor from Boeing, the 2707, in 1971, before it built any prototypes. What I’d like to talk about today is a renewed enthusiasm for supersonic passenger aircraft, and what’s changed that might make supersonic transport a viable market, today. — In the United States, commercial aircraft are not allowed to fly at supersonic speeds. This is because the sonic booms generated by supersonic flight, which are shockwaves that work a bit like the crack of a bullwhip or the firing of a bullet, but much, much larger, can set off alarms, rattle or shatter windows, and generally create all sorts of chaos on the ground, even in areas not directly under the aircraft that’s breaking the sound barrier. This was true even during the heyday of the Concorde: the craft was only allowed to travel at supersonic speeds over the ocean, because doing so over populated areas was such a pain, and in some cases, a danger. Sonic booms aren’t the only reason supersonic aircraft like the Concorde failed to establish a long-term presence in the airline industry, but they’re a big part of it. It’s just really difficult to work around that kind of persistent issue. This is why a new experimental project by NASA, the X-59 Quesst, with two-s’s, Quesst standing for Quiet SuperSonic Technology, is garnering so much attention. Built by Lockheed Martin, the X-59 is said to dramatically reduce the scale of sonic booms, instead producing what’s been described as a sonic thump, its long, slender nose breaking up the pressure waves that otherwise build up and create that much larger, more impactful shock wave boom, and its engine is on top of the plane rather than underneath it, a design choice that sends the majority of remaining shock wave impacts upward toward the sky, rather than down toward the ground. The X-59 is still just an experimental jet. It’s a single-seater, it’s about twice as long as an F-16 fighter jet, and it can cruise at around 925 miles per hours, which is Mach 1.4. It’s hoped that this new design will allow for the creation of future supersonic jetliners, though, as being able to traverse oceans twice as fast would bring massive economic benefits, in terms of shipping people, but also all kinds of goods. Being able to use these aircraft fully, at their full speed, over land and to and from any airport, would likewise make them more versatile and introduce new benefits and, hopefully, favorable economics. Worth noting here is that this jet is a descendent of that first Bell X-1 plane that broke the sound barrier in 1947; NASA’s X-planes are innovative models meant to push the boundaries of what’s currently possible, and the X-59 is just a more modern version of that initial X-1 conception in many ways. That said, the X-59 has only been successfully flown at low speeds and altitudes at this point. It got a lot of press at the end of October 2025 for successfully completing its first flight, which shows it can fly and land, which is good. But its inaugural flight stuck with a low altitude and just 240 miles per hour; really slow for a jet, and too low for a commercial airliner. The folks behind this project have also said that while they have every reason to believe this design will both work and create a far less impactful sonic boom, they don’t yet know if that boom will actually be tolerable for people on the ground. Simulating such things is different from the experience of them, and they won’t know until they power the thing all the way up and have it break the sound barrier whether the sonic thump will be barely noticeable and tolerable for folks near airports and flight paths, or if it will be better, but still not good enough to make this a viable alternative to existing jets. There are other entities working on similar things right now, including a company called Boom Supersonic that has already flown a piloted demonstration aircraft, the XB-1, at supersonic speeds—Mac 1.122, which is about 750 mph—at an altitude of over 35,000 feet; the first time a non-government-affiliated aircraft has done so. That was back in March of 2024, and the company plans to build a commercial supersonic aircraft that will carry between 64 and 80 passengers at Mach 1.7, on hundreds of global routes; they say they already have a large number of orders for this passenger aircraft they intend to build, and they say to begin with, they’ll be able to produce 66 of them per year from their factory in North Carolina. They say that they’ll have the first full-scale prototype of that passenger aircraft, called the Overture, in 2027, and they’re aiming to put that craft into service beginning in 2029 or 2030. They’re not the only private compan

04/11/2025

•

15min
21/10/2025

Circular Finance

This week we talk about entanglements, monopolies, and illusory money. We also discuss electrification, LLMs, and data centers. Recommended Book: The Extinction of Experience by Christine Rosen Transcript One of the big claims about artificial intelligence technologies, including but not limited to LLM-based generative AI tech, like ChatGPT, Claude, and Gemini, is that they will serve as universal amplifiers. Electricity is another universal amplifier, in that electrifying systems allows you to get a lot more from pretty much every single thing you do, while also allowing for the creation of entirely new systems. Cooking things in the kitchen? Much easier with electricity. Producing things on an assembly line? The introduction of electricity allows you to introduce all sorts of robotics, measuring tools, and safety measures that would not have otherwise been available, and all of these things make the entire process safer, cheaper, and a heck of a lot more effective and efficient. The prime argument behind many sky-high AI company valuations, then, is that if these things evolve in the way they could evolve, becoming increasingly capable and versatile and cheap, cooking could become even easier, manufacturing could become still faster, cheaper, and safer, and every other aspect of society and the economy would see similar gains. If you’re the people making AI, if you own these tools, or a share of the income derived from them, that’s a potentially huge pot of money: a big return on your investment. People make fortunes off far more focused, less-impactful companies and technologies all the time, and being able to create the next big thing in not just one space, but every space? Every aspect of everything, potentially? That’s like owning a share of electricity, and making money every time anyone uses electricity for anything. Through that lens, the big boom in both use of and investment in AI technologies maybe shouldn’t be so surprising. This represents a potentially generational sea-change in how everything works, what the economy looks like, maybe even how governments are run, militaries fight, and so on. If you can throw money into the mix, why wouldn’t you? And if that’s the case, the billions upon billions of dollars sloshing around in this corner of the tech world make a lot of sense; it may be curious that there’s not even more money being invested. Belief in that promise is not universal, however. A lot of people see these technologies not as the next electricity, but maybe the next smartphone, or perhaps the next SUV. Smartphones changed a whole lot about society too, but they’re hardly the same groundbreaking, omni-powerful upgrade that electricity represents. SUVs, too, flogged sales for flailing car companies, boosting their revenues at a moment in which they desperately needed to sell more vehicles to survive. But they were just another, more popular model of what already came before. There’s a chance AI will be similar to that: better software than came before, for some people’s use-cases—but not revolutionary, not groundbreaking even on the scale of pocketable phone-computers. What I’d like to talk about today are the peculiar economics that seem to be playing a role in the AI boom, and why many analysts and financial experts are eyeballing these economics warily, worrying about what they maybe represent, and possibly portend. — The term ‘exuberance,’ in the context of markets, refers to an excitement among investors—sometimes professional investors, sometimes casual investors, sometimes both—about a particular company, technology, or financial product type. The surge in interest and investment in cryptoassets during the height of the COVID-19 pandemic, for instance, including offshoot products like NFTs, was seemingly caused by a period of exuberance, sparked by the novelty of the product, the riches a few lucky insiders made off these products, and the desire by many people—pros and consumer-grade investors—to get in on that action, at a moment in which there wasn’t as much to do in the world as usual. Likewise, the gobs of money plowed into early internet companies, and the money thrown at companies laying fiberoptic cable for the presumed boom in internet customers, were, in retrospect, at least partly the consequence of irrational exuberance. In some cases these investors were just too early, as was the case with those cable-laying companies—the majority of them going out of business after blowing through a spectacular amount of money in a short period of time, and not finding enough paying customers to fund all that expansion—in others it was the result of sky-high valuations that were based on little beyond the exuberance of investors who probably should have known better, but who couldn’t get past their fear of missing out on the next big thing. In that latter case, that flow of money into early dotcom startups did fund a few winners that survived the eventual bursting of that bubble, but the majority of companies tagged with those massive valuations went out of business in part because their valuations were based in part on optimism, hot air, and illusory financials. Which is to say, their financials were based on a lot of money being added to their account sheets and tallied in the places investors would see those numbers, but the numbers didn’t mean what most people thought they meant. A company could receive tens of millions of dollars in orders, for instance, but that money and those orders might never be received and fulfilled, or that money might be mostly illusory: maybe it was borrowed from another company to spend on advertising, and that money would then go right back out the door, to the company from which it was borrowed, to pay for their ad services. That kind of arrangement could be beneficial, as the company doing the borrowing might give up a relatively small number of shares in exchange for money, which looks good on its balance sheet, especially if the money is given at a high valuation, even if that money was mostly just a loan from a company providing ad services, with the full knowledge that money would then be spent on their own ad services. And the ad company giving the money could usually afford to buy in at a high valuation, because it knows it will get that money right back, and when it does, it will get to record that money as income on its own balance sheets. So Company A gets millions of dollars from Company B, that money is then paid to Company B for some type of service, and both companies get to record favorable figures on their accounting sheets, as if real sales took place and real outside money changed hands, despite it being a circular move, with very little or no actual value being created. These sorts of relationships are also often good for investors in companies that do this sort of thing, because it makes their investments, the companies they’ve bought into, look even more valuable. Check it out, Company A, which I own shares in, is worth more than it was last month because of all the business it’s conducting, and because this other company bought into it at a higher price per share than I paid! Even though that increase in valuation is predicated on circular financing, the numbers still go up, and they go up for everyone involved, so there’s little reason to crack down on this not illegal, but shady behavior, and even less reason to want anyone else to know about it, because then they might not add their own money to the circular money-cycling, number-increasing machine. The major concern amongst some analysts right now is that the AI boom, especially in the United States, might be essentially this kind of circular cycle, but much larger than previous versions of the same. In the US right now, investment in AI infrastructure like data centers accounts for a huge portion of overall growth—the numbers vary, depending on who you ask and what numbers they look at, but some say that about 90% of total US economic growth, and around 80% of US stock market growth, are predicated on these sorts of investments this past year. Without these investments, the US economy would be basically flat, or worse, and the US stock market would be flailing as well. This situation isn’t ideal whatever the specifics, as too much reliance on just one industry, or one small collection of industries dominated by just a handful of companies and their investors, makes for a precarious financial foundation. If anything goes wrong with just one company, the whole house of cards could collapse. And if anything goes wrong with the industry, things could get even worse, and fast. All that investment, all that construction, all those employees and all that money sloshing around could disappear, could stop being spent, could make all those numbers fall and fall and fall more or less overnight. If this industry is in fact in a bubble, and if it’s being propped up by this kind of circular financing, where companies are fluffing up their own and each other’s accounting books by rotating the same bundle of money and on-paper money from company to company to company, that would portend pretty bad things for the US economy and market, if anyone involved stumbles, even just a little. This is why recent deals between the biggest players in this space are raising so many eyebrows, and causing so much sweat to bead on so many foreheads. In September of 2025, ChatGPT-maker OpenAI announced it had formalized a $100 billion investment deal with AI chipmaker Nvidia, the latter expanding on its existing investment in the former. In October, OpenAI announced it was purchasing billions of dollars worth of AI hardware from Nvidia-rival AMD, and that it’s taking a 10% stake in the company. Microsoft is already heavily invested in OpenAI, to the tune of $13 billion; it takes 49% of OpenAI’s profits, and gets more than that until its

21/10/2025

•

16min
23/12/2025

Wealth-Adjusted Life Satisfaction

More information about Brain Lenses at brainlenses.com. Paid BL supporters receive an additional episode of the show each week. Read the written version of this episode: This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit brainlenses.substack.com/subscribe

23/12/2025

•

3min
09/12/2025

Digital Asset Markets

This week we talk about in-game skins, investment portfolios, and Counter-Strike 2. We also discuss ebooks, Steam, and digital licenses. Recommended Book: Apple in China by Patrick McGee Transcript Almost always, if you buy an ebook or game or movie or music album online, you’re not buying that ebook, or that game, or whatever else—you’re buying a license that allows you access it, often on a specified device or in a specified way, and almost always in a non-transferrable, non-permanent manner. This distinction doesn’t matter much to most of us most of the time. If I buy an ebook, chances are I just want to read that ebook on the device I used to buy it, or the kindle attached to my Amazon or other digital book service account. So I buy the book, read it on my ebook reader or phone, and that’s that; same general experience I would have with a paperback or hardback book. This difference becomes more evident when you think about what happens to the book after you read it, though. If I own a hard-copy, physical book, I can resell it. I can donate it. I can put it in a Little Free Library somewhere in my neighborhood, or give it to a friend who I think will enjoy it. I can pick it up off my shelf later and read the exact same book I read years before. Via whichever mechanism I choose, I’m either holding onto that exact book for later, or I’m transferring ownership of that book, that artifact that contains words and/or images that can now be used, read, whatever by that second owner. And they can go on to do the same: handing it off to a friend, selling it on ebay, or putting it on a shelf for later reference. Often the convenience and immediacy of electronic books makes this distinction a non-issue for those who enjoy them. I can buy an ebook from Amazon or Bookshop.org and that thing is on my device within seconds, giving me access to the story or information that’s the main, valuable component of a book for most of us, without any delay, without having to drive to a bookstore or wait for it to arrive in the mail. That’s a pretty compelling offer. This distinction becomes more pressing, however, if I decide I want to go back and read an ebook I bought years ago, later, only to find that the license has changed and maybe that book is no longer accessible via the marketplace where I purchased it. If that happens, I no longer have access to the book, and there’s no recourse for this absence—I agreed to this possibility when I “bought” the book, based on the user agreement I clicked ‘OK’ or ‘I agree’ on when I signed up for Amazon or whichever service I paid for that book-access. It also becomes more pressing if, as has happened many times over the past few decades, the publisher or some other entity with control over these book assets decides to change them. A few years ago, for instance, British versions of Roald Dalh’s ‘Matilda’ were edited to remove references to Joseph Conrad, who has in recent times been criticized for his antisemitism and racist themes in his writing. Some of RL Stine’s Goosebumps books were edited to remove references to crushes schoolgirls had on their headmaster, and descriptions of an overweight character that were, in retrospect, determined to be offensive. And various racial and ethnic slurs were edited out of some of Agatha Christie’s works around the same time. Almost always, these changes aren’t announced by the publishers who own the rights to these books, and they’re typically only discovered by eagle-eyed readers who note that, for instance, the publishers decided to change the time period in which something occurred, which apparently happened in one of Stine’s works, without obvious purpose. This also frequently happens without the author being notified, as was the case with Stine and the edits made to his books. The publishers themselves, when asked directly about these changes, often remain silent on the matter. What I’d like to talk about today is another angle of this distinction between physically owned media and digital, licensed versions of the same, and the at times large sums of money that can be gained or lost based on the decisions of the companies that control these licensed assets. — Counter-Strike 2 is a first-person shooter game that’s free-to-play, was released in 2023, and was developed by a company called Valve. Valve has developed all sorts of games over the years, including the Counter-Strike, Half-Life, DOTA, and Portal games, but they’re probably best known for their Steam software distribution platform. Steam allows customers to buy all sorts of software, but mostly games through an interface that also provides chat services and community forums. But the primary utility of this platform is that it’s a marketplace for buying and selling games, and it has match-making features for online multiplayer games, serves as a sort of library for gamers, so all their games are launchable from one place, and it serves as a digital rights management hub, which basically means it helps game companies ensure users aren’t playing with pirated software—if you want to use steam to store and launch your games, they have to be legit, purchased games, not pirated ones. As of early 2025, it was estimated that Steam claimed somewhere between 75-80% of the PC gaming market, compared to competitors like the Epic Game Store, which was founded by the folks behind the wildly successful game, Fortnite, which can only claim something like 5%. And Counter-Strike is one of Valve’s, and Steam’s crown jewels. It’s a free-to-play game that was originally developed as a mod, a free add-on to another game Valve owns called Half-Life, but Valve bought up the rights to that mod and developed it into its own thing, releasing the initial entry in the series in 2000, several main-series games after that in subsequent years, and then Counter-Strike 2 came out in 2023, to much acclaim and fanfare. Counter-Strike 2 often has around a million players online, playing the game at any given moment, and its tournaments can attract closer to 1.5 million. As of early 2024, it was estimated that Counter-Strike 2 pulled in around a billion dollars a year for Valve, primarily via what are called Case Keys, which allow players to open in-game boxes, each key selling for $2.50. Valve also takes a 15% cut of all player-to-player sales of items conducted on the Steam Community Market, which is a secure ebay- or Amazon-like component of their platform where players can sell digital items from the game, which are primarily aesthetic add-ons, like skins for weapons, stickers, and clothing—things that allow players to look different in the game, as opposed to things that allow them to perform better, which would give players who spent the most money an unfair advantage and thus make the game less competitive and fun. Because this is a free game, though, and by many estimates a really balance and well-made one, a lot of people play it, and a lot of people want to customize the look of their in-game avatar. So being able to open in-game boxes that contain loot, and being able to buy and sell said loot on the Steam Community Market, has led to a rich secondary economy that makes that component of the game more interesting for players, while also earning Valve a whole lot of money on the backend for those keys and that cut of sales between players. In late-October of 2025, Valve announced a change in the rules for Counter-Strike 2, now allowing players to trade-up more item types, including previously un-trade-up-able items like gloves and knives, into higher-grade versions of the same. So common items could be bundled together and traded in for less common items, and those less common items could be bundled together and traded up for rare ones. This seems like a small move from the outside, but it roiled the CS2 in-game economy, by some estimates causing upwards of $2 billion to basically disappear overnight, because rare gloves and knives were at times valued at as much as $1.5 million; again, these are just aesthetic skins that change the look of a player’s avatar or weapons, but there’s enough demand for these things that some people are willing to pay that much for ultra-rare and unique glove and knife skins. Because of that demand, some players had taken to spending real money on these ultra-rare items, treating their in-game portfolios of skins as something like an investment portfolio. If you can buy an ultra-rare glove skin for $40,000 and maybe sell it later for twice that, that might seem like a really good investment, despite how strange it may seem to those not involved in this corner of the gaming world to spend $40,000 on what’s basically just some code in a machine that tells the game that the gloves on your avatar will look a certain way. This change, then, made those rarer gloves and knives, which were previously unattainable except by lottery-like chance, a lot more common, because people could trade up for them, increasing their chances of getting the ultra-rare stuff. The market was quickly flooded with more of these things, and about half the value of rare CS2 skins disappeared, initially knocking about $6 billion of total value from the market before stabilizing to around $1.5-2 billion. Volatility in this market continues, and people who invested a lot of money, sometimes their life savings, and sometimes millions of dollars into CS2 in-game skins, have been looking into potential legal recourse, though without much luck; Valve’s user agreements make very clear that players don’t own any of this stuff, and as a result, Valve can manipulate the market however they like, whenever they like. Just like with ebooks and movies we “buy” from Amazon and other services, then, these in-game assets are licensed to us, not sold. We may, at times, have a means of putting our license to some of these things on a secondary mar

09/12/2025

•

14min
07/10/2025

Gamewashing

This week we talk about Electronic Arts, 3DO, and the Saudi Arabian Public Investment Fund. We also discuss Jared Kushner, leveraged buyouts, and loot boxes. Recommended Book: Bandwidth by Dan Caruso Transcript Electronic Arts, often shorthanded as EA, was founded in 1982 in California by a former Apple employee named Trip Hawkins, who also went on to found the ill-fated 3DO company, which made video game hardware, and the somewhat more prolific, but also ultimately ill-fated casual game developer Digital Chocolate. EA, though, has been an absolutely astounding success. It’s business model was predicated on the premise of selling video games directly to retailers, rather than going through intermediaries. This allowed them to gain more market share than their competitors right off the bat, and it helped them glean higher margins than their competitors from each direct sale, too. EA also established an early reputation for treating its developers really well. They were the first gaming company to feature their developers in advertising and to give them platforms, promoting them as video game artists, basically, and it shared the profits netted from those direct sales with these develops—which in turn meant all the best developers really wanted to work for EA, which led to a beneficial cycle where they created better and better, and more and more financially successful games. In the late-80s, they started deviating from this model somewhat, scooping up a collection of successful independent game development studios and deviating, at times, from the creative lead’s vision when releasing their games. They also refocused a fair bit of their resources on franchises, like the immensely successful, as it turned out, Madden NFL series, and they branched out into producing games for the console market, including the still-new Nintendo Entertainment System, in 1990. That same year, EA went public on the NASDAQ, the company got new leadership when Hawkins decided to refocus on his far less successful 3DO hardware startup, and in an interesting twist, the arrival of the Sony Playstation in North America caused EA to drop support for 3DO hardware in the mid-90s so it could refocus on Playstation games, which were a lot more lucrative. By the mid-90s, EA had an astonishingly large and successful software library, including franchises like the aforementioned Madden games and the FIFA soccer games, but also celebrity-tied games like Shaq Fu, and military shooters like Jungle and Urban Strike. By the early-2000s, EA was making exclusive licensing deals with the NFL and ESPN, in order to stave off newfound sports game competitors, and it was the only video game company to consistently make a profit, most others experiencing feast and famine cycles, with periodic wins, but a whole lot of losses they had to cover with the profits from those wins. EA, in contrast, had a reliable stable of profit-sources, and it thus had a whole lot of leverage in terms of attracting and retaining talent, but also getting big names and brands on board, for collaborative projects. What I’d like to talk about today is what happened to EA during and following the 2008 economic crisis, and how and why it recently became an acquisition target for Saudi Arabia. — In 2008, when the global economy was collapsing, EA suffered a bad holiday sales season and fired 1,100 employees and closed 12 of their facilities early the following year. Later in 2009, the company announced the firing of another 1,500 employees, which was about 17% of their total workforce at the time, and in 2010 they acquired a gaming company that focused on mobile games, which were becoming increasingly popular, now that many people had touch-capable smartphones, which brought hot new franchises like Angry Birds under their brand umbrella. On the strength of that acquisition and all those downsizings, in early 2011, EA announced that it hit $3.8 billion in revenue in the financial year for the first time, and in early 2012, it announced it surpassed $1 billion in digital revenue during the previous year, which was a huge figure that early in the digital media landscape. It used some of those profits to scoop up another mobile-first gaming company, adding properties like Plants vs Zombies and Peggle to their library. EA completed another mass-firing in 2013, dismissing 10% of their employees under what they called a reorganization, around the same time they announced an exclusive license with Disney that would allow them to develop Star Wars games. Their stock value boomed in the following years, as a result of those cost-savings measures, and those new relationships, and emboldened by record-high stock valuations, in the mid-20-teens, the company started releasing big-name games, like Star Wars Battlefront 2, with random-content loot boxes and other sorts of microtransactions. This did not go over well with players, who decried these in-game purchasing options as ‘pay to win’ mechanics, as players could pay more money to get better characters and equipment, and a lot of the content, even after paying for the expensive games, was still locked behind paywalls, requiring more payments to unlock that content. A bunch of gaming journalists cried foul on this shift as the game careened toward its full release, as did a whole lot of early players, and Disney complained, too, so by the time it hit shelves, the game’s loot system was substantially changed, but that whole controversy spooked investors, and led to an 8.5% stock value drop in just a single month, knocking $3.1 billion from the company’s valuation. As a result of that controversy, EA also became the face for a larger legal and legislative debate about in-game purchases and how it’s kinda sorta like gambling, from that point forward. Soon after, EA experienced a series of bad quarters, including a huge drop of 13.3% to its valuation when a major entry in one of their larger franchises, Battlefield V, was released late, and received very mixed reviews when it was released, which led to a million fewer sold copies than anticipated. The game was also lagging in terms of gameplay behind smaller, nimbler competitors, including then-burgeoning Fortnite. The company saw an overall boost with the surprise success of Apex Legends, and the COVID-19 pandemic boosted sales dramatically for a while, since everyone was staying home, which allowed EA to gobble up a few more competing companies with successful franchises, and they knocked out a few more successful Star Wars games, as well. In early 2021, Saudi Arabia’s public investment funds bought 7.4 million shares of EA for about $1.1 billion, which flew under the radar for most gamers, but that’ll be important in a moment. Later that year, the company experienced a massive hack, a lot of its data, including the source code for games, stolen and sold on the dark web. EA bought some more competitors, but word on the street in 2022 was the the higher ups at EA were quietly shopping the company around, themselves looking to be acquired by a larger entity, on the scale of Apple or Disney. In early 2023, the company announced more mass-layoffs and launched another internal reorganization. It gutted several of its most popular gaming sub-brands, including BioWare, it cancelled an upcoming Star Wars game, and it announced that it would be shifting away from licensing agreements and refocusing on EA-owned IP. The pattern of layoffs leading to better financial fortunes didn’t pay off this time, though. In early 2025, EA divulged that it expected to underperform in the coming year, several of its big-name titles not doing as well as expected; the company cast blame on the market, but players and journalists pointed at the company’s gutting of its big-name studios, and the firing of many of its veteran developers to explain the reduced sales. EA had another mass-firing in April of this year, and followed by another in May, which paralleled an announcement that they would no longer be moving forward with a big, planned Black Panther game. In late September of 2025, EA announced that it had reached a deal, worth $55 billion, to go private, no longer selling shares on the stock market, with the financial assistance of a group of investors, which included Affinity Partners, which is led by Jared Kushner, US President Trump’s son-in-law, Silver Lake, which is a US-based private equity firm that helps make these sorts of big sales happen, and the aforementioned Saudi Arabian Public Investment Fund. This deal isn’t done yet, it still needs to get regulatory approval and a successful vote by stockholders, but it seems likely to go through, since the US regulatory environment is pretty lax at the moment, and because Kushner is involved, it’s unlikely President Trump will take a personal disliking to it. But the big story here seems to be that Saudi Arabia is buying up not just a video game company, but one of the biggest and most successful video gaming companies in the world, which, although it’s lost a lot of fan-credibility over the years, still owns some massively influential intellectual property and has just a stunning number of relationships and connections throughout the media world, alongside its huge valuation. If the sale does go through, and we should know for sure by sometime around June 2026, it would be the largest-ever leveraged buyout, which means the purchase was completed by using borrowed money that was borrowed against the asset being purchased; so those investors have taken out debt against EA itself, which is an increasingly common means of buying a large asset on the cheap, but it also typically burdens that asset with a simply astounding amount of debt which must then be recouped, often by selling off undervalued assets. When this happens to a newspaper, for instance, the buyer will often sell off the paper’s real estate and fire all their employees, t

07/10/2025

•

18min
25/11/2025

Thorium Reactors

This week we talk about radioactive waste, neutrons, and burn while breeding cycles. We also discuss dry casks, radioactive decay, and uranium. Recommended Book: Breakneck by Dan Wang Transcript Radioactive waste, often called nuclear waste, typically falls into one of three categories: low-level waste that contains a small amount of radioactivity that will last a very short time—this is stuff like clothes or tools or rags that have been contaminated—intermediate-level waste, which has been contaminated enough that it requires shielding, and high-level waste, which is very radioactive material that creates a bunch of heat because of all the radioactive decay, so it requires both shield and cooling. Some types of radioactive waste, particularly spent fuel of the kind used in nuclear power plants, can be reprocessed, which means separating it into other types of useful products, including another type of mixed nuclear fuel that can be used in lieu of uranium, though generally not economically unless uranium supplies are low. About a third of all spent nuclear fuel has already been reprocessed in some way. About 4% of even the recyclable stuff, though, doesn’t have that kind of second-life purpose, and that, combined with the medium- and long-lived waste that is quite dangerous to have just sitting around, has to be stored somehow, shielded and maybe cooled, and in some cases for a very long time: some especially long-lived fission products have half-lives that stretch into the hundreds of thousands or millions of years, which means they will be radioactive deep into the future, many times longer than humans have existed as a species. According to the International Atomic Energy Agency, something like 490,000 metric tons of radioactive spent fuel is currently being stored, on a temporary basis, at hundreds of specialized sites around the world. The majority of this radioactive waste is stored in pools of spent fuel water, cooled in that water somewhere near the nuclear reactors where the waste originated. Other waste has been relocated into what’re called dry casks, which are big, barrel-like containers made of several layers of steel, concrete, and other materials, which surround a canister that holds the waste, and the canister is itself surrounded by inert gas. These casks hold and cool waste using natural air convection, so they don’t require any kind of external power or water sources, while other solutions, including storage in water, sometimes does—and often the fuel is initially stored in pools, and is then moved to casks for longer-term storage. Most of the radioactive waste produced today comes in the form of spend fuel from nuclear reactors, which are typically small ceramic pellets made of low-enriched uranium oxide. These pellets are stacked on top of each other and encased in metal, and that creates what’s called a fuel rod. In the US, alone, about 2,000 metric tons of spent nuclear fuel is created each year, which is just shy of half an olympic sized swimming pool in terms of volume, and in many countries, the non-reuseable stuff is eventually buried, near the surface for the low- to intermediate-level waste, and deeper for high-level waste—deeper, in this context, meaning something like 200-1000 m, which is about 650-3300 feet, beneath the surface. The goal of such burying is to prevent potential leakage that might impact life on the surface, while also taking advantage of the inherent stability and cooler nature of underground spaces which are chosen for their isolation, natural barriers, and water impermeability, and which are also often reinforced with human-made supports and security, blocking everything off and protecting the surrounding area so nothing will access these spaces far into the future, and so that they won’t be broken open by future glaciation or other large-scale impacts, either. What I’d like to talk about today is another potential use and way of dealing with this type of waste, and why a recent, related development in China is being heralded as such a big deal. — An experimental nuclear reactor was built in the Gobi Desert by the Chinese Academy of Sciences Shanghai Institute of Applied Physics, and back in 2023 the group achieved its first criticality, got started up, basically, and it has been generating heat through nuclear fission ever since. What that means is that the nuclear reactor did what a nuclear reactor is supposed to do. Most such reactors exist to generate heat, which then creates steam and spins turbines, which generates electricity. What’s special about this reactor, though, is that it is a thorium molten salt reactor, which means it uses thorium instead of uranium as a fuel source, and the thorium is processed into uranium as part of the energy-making process, because thorium only contains trace amounts of fissile material, which isn’t enough to get a power-generating, nuclear chain reaction going. This reactor was able to successfully perform what’s called in-core thorium-to-uranium conversion, which allows the operators to use thorium as fuel, and have that thorium converted into uranium, which is sufficiently fissile to produce nuclear power, inside the core of the reactor. This is an incredibly fiddly process, and requires that the thorium-232 used as fuel absorb a neutron, which turns it into thorium-233. Thorium-233 then decays into protactinium-233, and that, in turn, decays into uranium-233—the fuel that powers the reactor. One innovation here is that this entire process happens inside the reactor, rather than occurring externally, which would require a bunch of supplementary infrastructure to handle fuel fabrication, increasing the amount of space and cost associated with the reactor. Those neutrons required to start the thorium conversion process are provided by small amounts of more fissile material, like enriched uranium-235 or plutonium-239, and the thorium is dissolved in a fluoride salt and becomes a molten mixture that allows it to absorb that necessary neutron, and go through that multi-step decay process, turning into uranium-233. That end-point uranium then releases energy through nuclear fission, and this initiates what’s called a burn while breeding cycle, which means it goes on to produce its own neutrons moving forward, which obviates the need for those other, far more fissile materials that were used to start the chain reaction. All of which makes this process a lot more fuel efficient than other options, dramatically reduces the amount of radioactive waste produced, and allows reactors that use it to operate a lot longer without needing to refuel, which also extends a reactor’s functional life. On that last point, many typical nuclear power plants built over the past handful of decades use pressurized water reactors which have to be periodically shut down so operators can replace spent fuel rods. This new method instead allows the fissile materials to continuously circulate, enabling on-the-fly refueling—so no shut-down, no interruption of operations necessary. This method also requires zero water, which could allow these reactors to be built in more and different locations, as conventional nuclear power plants have typically been built near large water sources, like oceans, because of their cooling needs. China initiated the program that led to the development of this experimental reactor back in 2011, in part because it has vast thorium reserves it wanted to tap in its pursuit of energy independence, and in part because this approach to nuclear energy should, in theory at least, allow plant operators to use existing, spent fuel rods as part of its process, which could be very economically interesting, as they could use the waste from their existing plants to help fuel these new plants, but also take such waste off other governments’ hands, maybe even be paid for it, because those other governments would then no longer need to store the stuff, and China could use it as cheap fuel; win win. Thinking further along, though, maybe the real killer application of this technology is that it allows for the dispersion of nuclear energy without the risk of nuclear weapons proliferation. The plants are smaller, they have a passive safety system that disallows the sorts of disasters that we saw in Chernobyl and Three-Mile Island—that sort of thing just can’t happen with this setup—and the fissile materials, aside from those starter materials used to get the initial cycle going, can’t be used to make nuclear weapons. Right now, there’s a fair amount of uranium on the market, but just like oil, that availability is cyclical and controlled by relatively few governments. In the future, that resource could become more scarce, and this reactor setup may become even more valuable as a result, because thorium is a lot cheaper and more abundant, and it’s less tightly controlled because it’s useless from a nuclear weapons standpoint. This is only the very first step on the way toward a potentially thorium-reactor dominated nuclear power industry, and the conversion rate on this experimental model was meager. That said, it is a big step in the right direction, and a solid proof-of-concept, showing that this type of reactor has promise and would probably work scaled-up, as well, and that means the 100MW demonstration reactor China is also building in the Gobi, hoping to prove the concept’s full value by 2035, stands a pretty decent chance of having a good showing. Show Notes https://www.deepisolation.com/about-nuclear-waste/where-is-nuclear-waste-now https://www.energy.gov/ne/articles/5-fast-facts-about-spent-nuclear-fuel https://www.energy.gov/ne/articles/3-advanced-reactor-systems-watch-2030 https://world-nuclear.org/information-library/nuclear-fuel-cycle/nuclear-waste/radioactive-wastes-myths-and-realities https://www.visualcapitalist.com/visualizing-all-the-nuclear-waste-in-the-world/ https://en

25/11/2025

•

13min
28/10/2025

Workplace Automation

This week we talk about robots, call center workers, and convenience stores. We also discuss investors, chatbots, and job markets. Recommended Book: The Fourth Consort by Edward Ashton Transcript Though LLM-based generative AI software, like ChatGPT, Gemini, and Claude, are becoming more and more powerful by the month, and offering newfangled functionality seemingly every day, it’s still anything but certain these tools, and the chatbots they power, will take gobs of jobs from human beings. The tale that’s being told by upper-management at a lot of companies makes it seem like this is inevitable, though there would seem to be market incentives for them to both talk and act like this is the case. Companies that make new, splashy investments in AI tech, or which make deals with big AI companies, purporting to further empower their offerings and to “rightsize” their staff as a consequence, tend to see small to moderate bumps in their stock price, and that’s good for the execs and other management in those companies, many of whom own a lot of stock, or have performance incentives related to the price of their stock built into their larger pay package. But often, not always, but quite a lot of the time, the increased effectiveness and efficiencies claimed by these higher-ups after they go on a firing spree and introduce new AI tools, seem to be at least partly, and in some cases mostly attributable to basically just threatening their staff with being fired in a difficult labor market. When Google executives lay off 5 or 10% of their staff on a given team, for instance, and then gently urge those who survived the cull to come to the office more frequently rather than working from home, and tell them that 60 hours a week is the sweet spot for achieving their productivity goals, that will tend to lead to greater outputs—at least for a while. Same as any other industry where blood has been drawn and a threat is made if people don’t live up to a casually stated standard presented by the person drawing that blood. Also worth mentioning here is that many of the people introducing these tools, both into their own companies and into the market as a whole, seem to think most jobs can be done by AI systems, but not theirs. Many executives have outright said that future businesses will have a small number of people managing a bunch of AI bots, and at least a few investors have said that they believe most jobs can be automated, but investing is too specialized and sophisticated, and will likely remain the domain of clever human beings like themselves. All of which gestures at what we’re seeing in labor markets around the globe right now, where demands for new hires are becoming more intense and a whole lot of low-level jobs in particular are disappearing entirely—though in most cases this is not because of AI, or not just, but instead because of automation more broadly; something that AI is contributing to, but something that is also a lot bigger than AI. And that’s what I’d like to talk about today. The rapid-speed deployment, in some industries and countries, at least, of automated systems, of robots, basically, and how this is likely to impact the already ailing labor markets in the places that are seeing the spearpoint of this deployment. — Chatbots are AI tools that are capable of taking input from users and responding with often quite human-sounding text, and increasingly, audio as well. These bots are the bane of some customers who are looking to speak to a human about some unique need or problem, but who are instead forced to run a gauntlet of AI-powered bots. The interaction often happens in the same little chat window through which they’ll eventually, if they say the right magic words, reach a human being capable of actually helping them. And like so many of the AI innovations that have been broadly deployed at this point, this is a solution that’s generally hated by customers, but lauded by the folks who run these companies, because it saves them a lot of money if they can hire fewer human beings to handle support tickets, even if those savings are the result of most people giving up before successfully navigating the AI maze and reaching a human customer support worker. In India right now, the thriving call center industry is seeing early signs of disruption from the same. IT training centers, in particular, are experimenting with using audio-capable AI chatbots instead of human employees, in part because demand is so high, but also, increasingly, because doing so is cheaper than hiring actual human beings to do the same work. One such company, LimeChat, recently said that it plans to cut its employee base by 80% in the near-future, and if that experiment is successful, this could ripple through India’s $283 billion IT sector, which accounts for 7.5% of India’s GDP. Hiring growth in this sector already collapsed in 2024 and 2025, and again, while this shift seems to be pretty good for the balance books of the companies doing less hiring and more firing as they deploy more AI systems, it’s very not good for the often younger people who take these jobs, specializing in call center IT work, only to find that the market no longer demands their skill sets. Along the same lines, but in a perhaps more surprising industry, some convenience stores in Japan are deploying robots to manage their back rooms, where the products that end up available out front are unloaded, tallied, and shelved. These robots, which are basically just arms on poles, sometimes attached to wheeled bases, for moving around, sometimes not, are operated by AI, but are also continuously monitored by human employees in the Philippines. Each worker, who can be paid a lot less than an entry-level, young Japanese person would expect to be paid, monitors about 50 machines at a time, and steps in, using virtual reality gear to control the robots, if one of them gets stuck or drops something; which apparently happens about 4% of the time. This is akin to offshoring of the kind we’ve seen since the early 2000s, when the dawn of technological globalization made China the factory of the world and everything shifted from a model of local production and the stockpiling of components, to a last-minute, supply-chain oriented model that allowed companies to move all their manufacturing and some of their services to wherever it could be done the cheapest. Many people and companies benefitted from this arbitrage to some degree, though many regions have dried up as a result of this shift, because, for instance, former company towns where cars were produced no longer have the resources to keep infrastructure from degrading, and no longer have enough jobs to keep young people from moving away; brain drain can become pretty intense when there’s no economic reason to stay. This reality is expected to become more widespread, even beyond former manufacturing hubs, because of the deployment of both AI systems, which can be subbed-in for many remote jobs, like call center work, programming, and the like, but also because of increasingly sophisticated and capable robots, which can do more automated work, which in turn allows them to be monitored, sometimes remotely, like those Japanese convenience store robots, for a fraction of the price of hiring a human being. This shift is expected to be especially harrowing for teens hoping to enter the labor market in entry-level jobs, as responsibilities like shelf-stocking and product scanning and the loading and unloading of materials are increasingly automatable, as robots capable of doing this work are developed and deployed, and perhaps even more importantly, as systems that augment that automatability are developed and deployed. In practice, that means coming up with shipping processes and other non-tangible systems that lean into the strengths of today’s automated systems, while reducing the impact of their weaknesses. Amazon is in prime position to do exactly this, as they’ve already done so much to rewire global shipping channels so that they can deliver products as rapidly as possible, to as many places as possible. As a result, they control many of the variables within these channels, which in turn means they can tweak them further, so that they’re optimized to work with Amazon’s specialized automated systems, rather than just human ones. The company has stated, in internal documents, that it plans to automate 75% of its total operations, and it currently has nearly 1.2 million employees. That’s triple what it employed in 2018, and it’s expected that the automated systems it has already and will soon deploy will allow it to hire 160,000 fewer people than planned by 2027. Even though the company expects to sell twice as many products by 2033, then, it expects to hire 600,000 fewer people by that same year. And it’s so confident in its ability to make this happen that it’s already making plans to rebuild its image in the aftermath of what’s expected to be a really difficult period of people hating it. It’s planning significant branding efforts, meant to help it seem like a good corporate citizens, including sponsored community events and big donations to children’s programs. It’s also intending to frame this shift as an evolution in which robots are amplifying the efforts of human employees. Rather than calling their automated systems robots, they might call them ‘cobots,’ for instance. Amazon has contended that the internal documents in which these plans were outlined, those documents acquired and reported upon by the New York Times, are incomplete and not an accurate representation of what Amazon plans, and they said those branding efforts are not a response to hate related to their automation efforts, they just like spending money on nice things for communities. The net-impact of existing efforts of this kind, though, is to deplete local job markets where thes

28/10/2025

•

16min

A portfolio of podcasts for curious, adventurous, passionate lifelong learners.

Site