Author name: Kelly Newman

Trump backs down a bit on auto industry tariffs—but only a bit

Cars, trump tariffs / Kelly Newman / April 29, 2025

President Donald Trump is set to ease up slightly on the automotive industry this week. After being warned that his trade war will result in hiked prices and fewer vehicles being built, government officials over the past two days have signaled that Trump will sign an executive order today that will mitigate some of the pain the 25 percent import tariffs will inflict.

Trump’s approach to tariffs has been nothing if not inconsistent. In this case, the White House is not dropping the 25 percent tariff on all imported vehicles, but the other tariffs imposed by the Trump administration—like the 25 percent tariff on steel and aluminum that went into effect in February—won’t stack up on top.

The potential for multiple tariffs to have an additive effect on prices could have seen new car prices soar in the coming weeks; now, they are likely to just rise a lot instead. According to The Wall Street Journal, the move will be retroactive, and automakers who have (for example) paid aluminum or steel tariffs on top of the car import tariff can seek a refund for the former.

A separate 25 percent tariff on imported car parts is set to go into effect on Saturday. These are still planned to go into effect, but importers will be able to apply for small reimbursements—up to 3.75 percent of the value of a US-made car for a year, then 2.5 percent of the car’s value for the second year.

How easy the government will make the reimbursement process, and where that money will come from, has yet to be made clear.

Trump backs down a bit on auto industry tariffs—but only a bit Read More »

Physics of the perfect cacio e pepe sauce

fluid dynamics, Food science, phase separation, Physics, rheology, Science / Kelly Newman / April 29, 2025

The trick: Add corn starch separately to make the sauce rather than using pasta water.

Cacio e pepe is an iconic pasta dish that can be frustratingly difficult to make Credit: Simone Frau

Nobody does pasta quite like the Italians, as anyone who has tasted an authentic “pasta alla cacio e pepe” can attest. It’s a simple dish: just tonnarelli pasta, pecorino cheese, and pepper. But its simplicity is deceptive. Cacio e pepe (“cheese and pepper”) is notoriously challenging to make because it’s so easy for the sauce to form unappetizing clumps with a texture more akin to stringy mozzarella rather than being smooth and creamy.

A team of Italian physicists has come to the rescue with a foolproof recipe based on their many scientific experiments, according to a new paper published in the journal Physics of Fluids. The trick: using corn starch for the cheese and pepper sauce instead of relying on however much starch leaches into the boiling water as the pasta is cooked.

“A true Italian grandmother or a skilled home chef from Rome would never need a scientific recipe for cacio e pepe, relying instead on instinct and years of experience,” the authors wrote. “For everyone else, this guide offers a practical way to master the dish. Preparing cacio e pepe successfully depends on getting the balance just right, particularly the ratio of starch to cheese. The concentration of starch plays a crucial role in keeping the sauce creamy and smooth, without clumps or separation.”

There has been a surprising amount of pasta-related physics research in recent years, particularly around spaghetti—the mechanics of slurping the pasta into one’s mouth, for instance, or spitting it out (aka, the “reverse spaghetti problem”). The most well-known is the question of how to get dry spaghetti strands to break neatly in two rather than three or more scattered pieces. French physicists successfully explained the dynamics in an Ig Nobel Prize-winning 2006 paper. They found that, counterintuitively, a dry spaghetti strand produces a “kick back” traveling wave as it breaks. This wave temporarily increases the curvature in other sections, leading to many more breaks.

In 2020, physicists provided an explanation for why a strand of spaghetti in a pot of boiling water will start to sag as it softens before sinking to the bottom of the pot and curling back on itself in a U shape. Physicists have also discovered a way to determine if one’s spaghetti is perfectly done by using a simple ruler (although one can always use the tried-and-true method of flinging a test strand against the wall). In 2021, inspired by flat-packed furniture, scientists came up with an ingenious solution to packaging differently shaped pastas: ship them in a flat 2D form that takes on the final 3D shape when cooked, thanks to carefully etched patterns in the pasta.

And earlier this year, physicists investigated how adding salt to a pasta pot to make it boil faster can leave a white ring on the bottom of the pot to identify factors leading to the perfect salt ring. They found that particles released from a smaller height fall faster and form a pattern with a clean central region. Those released from a greater height take longer to fall to the bottom, and the cloud of particles expands radially until the particles are far enough apart not to be influenced by the wakes of neighboring particles, such that they no longer form a cloud. In the latter case, you end up with a homogeneous salt ring deposit.

Going through a phase (separation)

ompare the effect of: water alone; pasta water that retains some starch (obtained by cooking 100 g of pasta in 1 liter of water); and pasta water “risottata” — Comparing the effect of water alone, pasta water that retains some starch, and pasta water “risottata.” Credit: G. Bartolucci et al., 2025

So it shouldn’t be the least bit surprising that physicists have now turned their attention to the problem of the perfect cacio e pepe sauce. The authors are well aware that they are treading on sacred ground for Italian traditionalists. “I hope that eight Italian authors is enough [to quell skepticism],” co-author Ivan Di Terlizzi of the Max Planck Institute for the Physics of Complex Systems told The New York Times back in January. (An earlier version of the paper was posted to the physics preprint arXiv in January, prompting that earlier coverage.)

Terlizzi and his fellow author are all living abroad and frequently meet for dinner. Cacio e pepe is among their favorite traditional dishes to make, and as physicists, they couldn’t help but want to learn more about the unique physics of the process, not to mention “the more practical aim to avoid wasting good pecorino,” said Terlizzi. They focused on the separation that often occurs when cheese and water are mixed, building on earlier culinary experiments.

As the pasta cooks in boiling water, the noodles release starch. Traditionally, the chef will extract part of the water and starch solution—which is cooled to a suitable temperature to avoid clumping as the cheese proteins “denaturate”—and mix it with the cheese to make the sauce, adding the pepper last, right before serving. But the authors note that temperature is not the only factor that can lead to this dreaded “mozzarella phase.”

According to the authors, if one tries to mix cheese and water without any starch, the clumping is more pronounced. There is less clumping with water containing a little starch, like water in which pasta has been cooked. And when one mixes the cheese with pasta water “risottata”—i.e., collected and heated in a pan so enough water evaporates that there is a higher concentration of starch—there is almost no clumping.

Effect of trisodium citrate on the stability of Cacio e pepe sauce. — Effect of trisodium citrate on the stability of cacio e pepe sauce. Credit: G. Bartolucci et al., 2025

So starch plays a crucial role in the process of making cacio e pepe. The authors devised a set of experiments to scientifically investigate the phase behavior of water, starch, and cheese mixed together in various concentrations and at different temperatures. They primarily used standard kitchen tools to make sure home cooks could recreate their results (although not every kitchen has a sous vide machine). This enabled them to devise a phase diagram of what happens to the sauce as the conditions change.

The authors found that the correct starch ratio is between 2 to 3 percent of the cheese weight. Below that, you get the clumping phase separation; above that, and the sauce “becomes stiff and unappetizing as it cools,” they wrote. Pasta water alone contains too little starch. Using pasta water “risottata” may concentrate the starch, but the chef has less control over the precise amount of starch. So the authors recommend simply dissolving 4 grams of powdered potato or corn starch in 40 grams of water, heating it gently until it thickens—a transition known as starch gelatinization—and combining that gel with the cheese. They also recommend toasting the black pepper briefly before adding it to the mixture to enhance its flavors and aromas.

They ran the same set of experiments using trisodium citrate as an alternative stabilizer, which is widely used in the food industry as an emulsifier—including in the production of processed cheese, since it enhances smoothness and prevents unwanted clumping, exactly the properties one desires for a perfect cacio e pepe sauce. The trisodium citrate at concentrations above 2 percent worked just as well at avoiding the mozzarella phase, “though at a cost of deviating from strict culinary tradition,” the authors concluded. “However, while the sauce stabilization is more efficient, we found the taste of the cheese to be slightly blunted, likely due to the basic properties of the salt.”

The team’s next research goal is to conduct similar experiments with making pasta alla gricia—basically the same as cacio e pepe, with the addition of guanciale (cured pork cheek). “This recipe seems to be easier to perform, and we don’t know exactly why,” said co-author Daniel Maria Busiello, Terlizzi’s colleague at the Dresden Max Planck Institute. “This is one idea we might explore in the future.”

DOI: Physics of Fluids, 2025. 10.1063/5.0255841 (About DOIs).

Jennifer is a senior writer at Ars Technica with a particular focus on where science meets culture, covering everything from physics and related interdisciplinary topics to her favorite films and TV series. Jennifer lives in Baltimore with her spouse, physicist Sean M. Carroll, and their two cats, Ariel and Caliban.

Physics of the perfect cacio e pepe sauce Read More »

Monty Python and the Holy Grail turns 50

culture, Entertainment, Features, film, film anniversaries, Monty Python, Monty Python and the Holy Grail / Kelly Newman / April 29, 2025

Ars staffers reflect upon the things they love most about this masterpiece of absurdist comedy.

Credit: EMI Films/Python (Monty) Pictures

Monty Python and the Holy Grail is widely considered to be among the best comedy films of all time, and it’s certainly one of the most quotable. This absurdist masterpiece sending up Arthurian legend turns 50 (!) this year.

It was partly Python member Terry Jones’ passion for the Middle Ages and Arthurian legend that inspired Holy Grail and its approach to comedy. (Jones even went on to direct a 2004 documentary, Medieval Lives.) The troupe members wrote several drafts beginning in 1973, and Jones and Terry Gilliam were co-directors—the first full-length feature for each, so filming was one long learning process. Reviews were mixed when Holy Grail was first released—much like they were for Young Frankenstein (1974), another comedic masterpiece—but audiences begged to differ. It was the top-grossing British film screened in the US in 1975. And its reputation has only grown over the ensuing decades.

The film’s broad cultural influence extends beyond the entertainment industry. Holy Grail has been the subject of multiple scholarly papers examining such topics as its effectiveness at teaching Arthurian literature or geometric thought and logic, the comedic techniques employed, and why the depiction of a killer rabbit is so fitting (killer rabbits frequently appear drawn in the margins of Gothic manuscripts). My personal favorite was a 2018 tongue-in-cheek paper on whether the Black Knight could have survived long enough to make good on his threat to bite King Arthur’s legs off (tl;dr: no).

So it’s not at all surprising that Monty Python and the Holy Grail proved to be equally influential and beloved by Ars staffers, several of whom offer their reminiscences below.

They were nerd-gassing before it was cool

The Monty Python troupe famously made Holy Grail on a shoestring budget—so much so that they couldn’t afford to have the knights ride actual horses. (There are only a couple of scenes featuring a horse, and apparently it’s the same horse.) Rather than throwing up their hands in resignation, that very real constraint fueled the Pythons’ creativity. The actors decided the knights would simply pretend to ride horses while their porters followed behind, banging halves of coconut shells together to mimic the sound of horses’ hooves—a time-honored Foley effect dating back to the early days of radio.

Being masters of absurdist humor, naturally, they had to call attention to it. Arthur and his trusty servant, Patsy (Gilliam), approach the castle of their first potential recruit. When Arthur informs the guards that they have “ridden the length and breadth of the land,” one of the guards isn’t having it. “What, ridden on a horse? You’re using coconuts! You’ve got two empty halves of coconut, and you’re bangin’ ’em together!”

That raises the obvious question: Where did they get the coconuts? What follows is one of the greatest examples of nerd-gassing yet to appear on film. Arthur claims he and Patsy found them, but the guard is incredulous since the coconut is tropical and England is a temperate zone. Arthur counters by invoking the example of migrating swallows. Coconuts do not migrate, but Arthur suggests they could be carried by swallows gripping a coconut by the husk.

The guard still isn’t having it. It’s a question of getting the weight ratios right, you see, to maintain air-speed velocity. Another guard gets involved, suggesting it might be possible with an African swallow, but that species is non-migratory. And so on. The two are still debating the issue as an exasperated Arthur rides off to find another recruit.

The best part? There’s a callback to that scene late in the film when the knights must answer three questions to cross the Bridge of Death or else be chucked into the Gorge of Eternal Peril. When it’s Arthur’s turn, the third question is “What is the air-speed velocity of an unladen swallow?” Arthur asks whether this is an African or a European swallow. This stumps the Bridgekeeper, who gets flung into the gorge. Sir Belvedere asks how Arthur came to know so much about swallows. Arthur replies, “Well, you have to know these things when you’re a king, you know.”

The plucky Black Knight (“It’s just a flesh wound!”) will always hold a special place in my heart, but that debate over air-speed velocities of laden versus unladen swallows encapsulates what makes Holy Grail a timeless masterpiece.

—Jennifer Ouellette

A bunny out for blood

“Oh, it’s just a harmless little bunny, isn’t it?”

Despite their appearances, rabbits aren’t always the most innocent-looking animals. Recent reports of rabbit strikes on airplanes are the latest examples of the mayhem these creatures of chaos can inflict on unsuspecting targets.

I learned that lesson a long time ago, though, thanks partly to my way-too-early viewings of the animated Watership Down and Monty Python and the Holy Grail. There I was, about 8 years old and absent of paternal accompaniment, watching previously cuddly creatures bloodying each other and severing the heads of King Arthur’s retinue. While Watership Down’s animal-on-animal violence might have been a bit scarring at that age, I enjoyed the slapstick humor of the Rabbit of Caerbannog scene (many of the jokes my colleagues highlight went over my head upon my initial viewing).

Despite being warned of the creature’s viciousness by Tim the Enchanter, the Knights of the Round Table dismiss the Merlin stand-in’s fear and charge the bloodthirsty creature. But the knights quickly realize they’re no match for the “bad-tempered rodent,” which zips around in the air, goes straight for the throat, and causes the surviving knights to run away in fear. If Arthur and his knights possessed any self-awareness, they might have learned a lesson about making assumptions about appearances.

But hopefully that’s a takeaway for viewers of 1970s British pop culture involving rabbits. Even cute bunnies, as sweet as they may seem initially, can be engines of destruction: “Death awaits you all with nasty, big, pointy teeth.”

—Jacob May

Can’t stop the music

The most memorable songs from Monty Python and the Holy Grail were penned by Neil Innes, who frequently collaborated with the troupe and appears in the film. His “Brave Sir Robin” amusingly parodied minstrel tales of valor by imagining all the torturous ways that one knight might die. Then there’s his “Knights of the Round Table,” the first musical number performed by the cast—if you don’t count the monk chants punctuated with slaps on the head with wooden planks. That song hilariously rouses not just wild dancing from knights but also claps from prisoners who otherwise dangle from cuffed wrists.

But while these songs have stuck in my head for decades, Monty Python’s Terry Jones once gave me a reason to focus on the canned music instead, and it weirdly changed the way I’ve watched the movie ever since.

Back in 2001, Jones told Billboard that an early screening for investors almost tanked the film. He claimed that after the first five minutes, the movie got no laughs whatsoever. For Jones, whose directorial debut could have died in that moment, the silence was unthinkable. “It can’t be that unfunny,” he told Billboard. “There must be something wrong.”

Jones soon decided that the soundtrack was the problem, immediately cutting the “wonderfully rich, atmospheric” songs penned by Innes that seemed to be “overpowering the funny bits” in favor of canned music.

Reading this prompted an immediate rewatch because I needed to know what the first bit was that failed to get a laugh from that fateful audience. It turned out to be the scene where King Arthur encounters peasants in a field who deny knowing that there even was a king. As usual, I was incapable of holding back a burst of laughter when one peasant woman grieves, “Well, I didn’t vote for you” while packing random clumps of mud into the field. It made me wonder if any song might have robbed me of that laugh, and that made me pay closer attention to how Jones flipped the script and somehow meticulously used the canned music to extract more laughs.

The canned music was licensed from a British sound library that helped the 1920s movie business evolve past silent films. They’re some of the earliest songs to summon emotion from viewers whose eyes were glued to a screen. In Monty Python and the Holy Grail, which features a naive King Arthur enduring his perilous journey on a wood stick horse, the canned music provides the most predictable soundtrack you could imagine that might score a child’s game of make-believe. It also plays the straight man by earnestly pulsing to convey deep trouble as knights approach the bridge of death or heavenly trumpeting the anticipated appearance of the Holy Grail.

It’s easy to watch the movie without noticing the canned music, as the colorful performances are Jones’ intended focus. Not relying on punchlines, the group couldn’t afford any nuance to be lost. But there is at least one moment where Jones obviously relies on the music to overwhelm the acting to compel a belly laugh. Just before “the most foul, cruel, bad-tempered rodent” appears, a quick surge of dramatic music that cuts out just as suddenly makes it all the more absurd when the threat emerges and appears to be an “ordinary rabbit.”

It’s during this scene, too, that King Arthur delivers a line that sums up how predictably odd but deceptively artful the movie’s use of canned music really is. When he meets Tim the Enchanter—who tries to warn the knights about the rabbit’s “pointy teeth” by evoking loud thunder rolls and waggling his fingers in front of his mouth—Arthur turns to the knights and says, “What an eccentric performance.”

—Ashley Belanger

Thank the “keg rock conclave”

I tried to make music a big part of my teenage identity because I didn’t have much else. I was a suburban kid with a B-minus/C-plus average, no real hobbies, sports, or extra-curriculars, plus a deeply held belief that Nine Inch Nails, the Beastie Boys, and Aphex Twin would never get their due as geniuses. Classic Rock, the stuff jocks listened to at parties and practice? That my dad sang along to after having a few? No thanks.

There were cultural heroes, there were musty, overwrought villains, and I knew the score. Or so I thought.

I don’t remember exactly where I found the little fact that scarred my oppositional ego forever. It might have been Spin magazine, a weekend MTV/VH1 feature, or that Rolling Stone book about the ’70s (I bought it for the punks, I swear). But at some point, I learned that a who’s-who of my era’s played-out bands—Led Zeppelin, Pink Floyd, even Jethro (freaking) Tull—personally funded one of my favorite subversive movies. Jimmy Page and Robert Plant, key members of the keg-rock conclave, attended the premiere.

It was such a small thing, but it raised such big, naive, adolescent questions. Somebody had to pay for Holy Grail—it didn’t just arrive as something passed between nerds? People who make things I might not enjoy could financially support things I do enjoy? There was a time when today’s overcelebrated dinosaurs were cool and hip in the subculture? I had common ground with David Gilmour?

Ever since, when a reference to Holy Grail is made, especially to how cheap it looks, I think about how I once learned that my beloved nerds (or theater kids) wouldn’t even have those coconut horses were it not for some decent-hearted jocks.

—Kevin Purdy

A masterpiece of absurdism

“I blow my nose at you, English pig-dog!” EMI Films/Python (Monty) Pictures

I was young enough that I’d never previously stayed awake until midnight on New Year’s Eve. My parents were off to a party, my younger brother was in bed, and my older sister had a neglectful attitude toward babysitting me. So I was parked in front of the TV when the local PBS station aired a double feature of The Yellow Submarine and The Holy Grail.

At the time, I probably would have said my mind was blown. In retrospect, I’d prefer to think that my mind was expanded.

For years, those films mostly existed as a source of one-line evocations of sketch comedy nirvana that I’d swap with my friends. (I’m not sure I’ve ever lacked a group of peers where a properly paced “With… a herring!” had meaning.) But over time, I’ve come to appreciate other ways that the films have stuck with me. I can’t say whether they set me on an aesthetic trajectory that has continued for decades or if they were just the first things to tickle some underlying tendencies that were lurking in my not-yet-fully-wired brain.

In either case, my brain has developed into a huge fan of absurdism, whether in sketch comedy, longer narratives like Arrested Development or the lyrics of Courtney Barnett. Or, let’s face it, any stream of consciousness lyrics I’ve been able to hunt down. But Monty Python remains a master of the form, and The Holy Grail’s conclusion in a knight bust remains one of its purest expressions.

A bit less obviously, both films are probably my first exposures to anti-plotting, where linearity and a sense of time were really besides the point. With some rare exceptions—the eating of Sir Robin’s minstrels, Ringo putting a hole in his pocket—the order of the scenes were completely irrelevant. Few of the incidents had much consequence for future scenes. Since I was unused to staying up past midnight at that age, I’d imagine the order of events was fuzzy already by the next day. By the time I was swapping one-line excerpts with friends, it was long gone. And it just didn’t matter.

In retrospect, I think that helped ready my brain for things like Catch-22 and its convoluted, looping, non-Euclidean plotting. The novel felt like a revelation when I first read it, but I’ve since realized it fits a bit more comfortably within a spectrum of works that play tricks with time and find clever connections among seemingly random events.

I’m not sure what possessed someone to place these two films together as appropriate New Year’s Eve programming. But I’d like to think it was more intentional than I had any reason to suspect at the time. And I feel like I owe them a debt.

—John Timmer

A delightful send-up of autocracy

King Arthur attempting to throttle a peasant in the field — “See the violence inherent in the system!” Credit: Python (Monty) Pictures

What an impossible task to pick just a single thing I love about this film! But if I had to choose one scene, it would be when a lost King Arthur comes across an old woman—but oops, it’s actually a man named Dennis—and ends up in a discussion about medieval politics. Arthur explains that he is king because the Lady of the Lake conferred the sword Excalibur on him, signifying that he should rule as king of the Britons by divine right.

To this, Dennis replies, “Strange women lying in ponds distributing swords is no basis for a system of government. Supreme executive power derives from a mandate from the masses, not from some farcical aquatic ceremony.”

Even though it was filmed half a century ago, the scene offers a delightful send-up of autocracy. And not to be too much of a downer here, but all of us living in the United States probably need to be reminded that living in an autocracy would suck for a lot of reasons. So let’s not do that.

—Eric Berger

Monty Python and the Holy Grail turns 50 Read More »

iOS and Android juice jacking defenses have been trivial to bypass for years

android, Apple, Biz & IT, choicejacking, Features, Google, iOS, juice jacking, juicejacking, Security, Tech / Kelly Newman / April 28, 2025

SON OF JUICE JACKING ARISES

New ChoiceJacking attack allows malicious chargers to steal data from phones.

Credit: Aurich Lawson | Getty Images

About a decade ago, Apple and Google started updating iOS and Android, respectively, to make them less susceptible to “juice jacking,” a form of attack that could surreptitiously steal data or execute malicious code when users plug their phones into special-purpose charging hardware. Now, researchers are revealing that, for years, the mitigations have suffered from a fundamental defect that has made them trivial to bypass.

“Juice jacking” was coined in a 2011 article on KrebsOnSecurity detailing an attack demonstrated at a Defcon security conference at the time. Juice jacking works by equipping a charger with hidden hardware that can access files and other internal resources of phones, in much the same way that a computer can when a user connects it to the phone.

An attacker would then make the chargers available in airports, shopping malls, or other public venues for use by people looking to recharge depleted batteries. While the charger was ostensibly only providing electricity to the phone, it was also secretly downloading files or running malicious code on the device behind the scenes. Starting in 2012, both Apple and Google tried to mitigate the threat by requiring users to click a confirmation button on their phones before a computer—or a computer masquerading as a charger—could access files or execute code on the phone.

The logic behind the mitigation was rooted in a key portion of the USB protocol that, in the parlance of the specification, dictates that a USB port can facilitate a “host” device or a “peripheral” device at any given time, but not both. In the context of phones, this meant they could either:

Host the device on the other end of the USB cord—for instance, if a user connects a thumb drive or keyboard. In this scenario, the phone is the host that has access to the internals of the drive, keyboard or other peripheral device.
Act as a peripheral device that’s hosted by a computer or malicious charger, which under the USB paradigm is a host that has system access to the phone.

An alarming state of USB security

Researchers at the Graz University of Technology in Austria recently made a discovery that completely undermines the premise behind the countermeasure: They’re rooted under the assumption that USB hosts can’t inject input that autonomously approves the confirmation prompt. Given the restriction against a USB device simultaneously acting as a host and peripheral, the premise seemed sound. The trust models built into both iOS and Android, however, present loopholes that can be exploited to defeat the protections. The researchers went on to devise ChoiceJacking, the first known attack to defeat juice-jacking mitigations.

“We observe that these mitigations assume that an attacker cannot inject input events while establishing a data connection,” the researchers wrote in a paper scheduled to be presented in August at the Usenix Security Symposium in Seattle. “However, we show that this assumption does not hold in practice.”

The researchers continued:

We present a platform-agnostic attack principle and three concrete attack techniques for Android and iOS that allow a malicious charger to autonomously spoof user input to enable its own data connection. Our evaluation using a custom cheap malicious charger design reveals an alarming state of USB security on mobile platforms. Despite vendor customizations in USB stacks, ChoiceJacking attacks gain access to sensitive user files (pictures, documents, app data) on all tested devices from 8 vendors including the top 6 by market share.

In response to the findings, Apple updated the confirmation dialogs in last month’s release of iOS/iPadOS 18.4 to require a user authentication in the form of a PIN or password. While the researchers were investigating their ChoiceJacking attacks last year, Google independently updated its confirmation with the release of version 15 in November. The researchers say the new mitigation works as expected on fully updated Apple and Android devices. Given the fragmentation of the Android ecosystem, however, many Android devices remain vulnerable.

All three of the ChoiceJacking techniques defeat the original Android juice-jacking mitigations. One of them also works against those defenses in Apple devices. In all three, the charger acts as a USB host to trigger the confirmation prompt on the targeted phone.

The attacks then exploit various weaknesses in the OS that allow the charger to autonomously inject “input events” that can enter text or click buttons presented in screen prompts as if the user had done so directly into the phone. In all three, the charger eventually gains two conceptual channels to the phone: (1) an input one allowing it to spoof user consent and (2) a file access connection that can steal files.

An illustration of ChoiceJacking attacks. (1) The victim device is attached to the malicious charger. (2) The charger establishes an extra input channel. (3) The charger initiates a data connection. User consent is needed to confirm it. (4) The charger uses the input channel to spoof user consent. Credit: Draschbacher et al.

It’s a keyboard, it’s a host, it’s both

In the ChoiceJacking variant that defeats both Apple- and Google-devised juice-jacking mitigations, the charger starts as a USB keyboard or a similar peripheral device. It sends keyboard input over USB that invokes simple key presses, such as arrow up or down, but also more complex key combinations that trigger settings or open a status bar.

The input establishes a Bluetooth connection to a second miniaturized keyboard hidden inside the malicious charger. The charger then uses the USB Power Delivery, a standard available in USB-C connectors that allows devices to either provide or receive power to or from the other device, depending on messages they exchange, a process known as the USB PD Data Role Swap.

A simulated ChoiceJacking charger. Bidirectional USB lines allow for data role swaps. Credit: Draschbacher et al.

With the charger now acting as a host, it triggers the file access consent dialog. At the same time, the charger still maintains its role as a peripheral device that acts as a Bluetooth keyboard that approves the file access consent dialog.

The full steps for the attack, provided in the Usenix paper, are:

1. The victim device is connected to the malicious charger. The device has its screen unlocked.

2. At a suitable moment, the charger performs a USB PD Data Role (DR) Swap. The mobile device now acts as a USB host, the charger acts as a USB input device.

3. The charger generates input to ensure that BT is enabled.

4. The charger navigates to the BT pairing screen in the system settings to make the mobile device discoverable.

5. The charger starts advertising as a BT input device.

6. By constantly scanning for newly discoverable Bluetooth devices, the charger identifies the BT device address of the mobile device and initiates pairing.

7. Through the USB input device, the charger accepts the Yes/No pairing dialog appearing on the mobile device. The Bluetooth input device is now connected.

8. The charger sends another USB PD DR Swap. It is now the USB host, and the mobile device is the USB device.

9. As the USB host, the charger initiates a data connection.

10. Through the Bluetooth input device, the charger confirms its own data connection on the mobile device.

This technique works against all but one of the 11 phone models tested, with the holdout being an Android device running the Vivo Funtouch OS, which doesn’t fully support the USB PD protocol. The attacks against the 10 remaining models take about 25 to 30 seconds to establish the Bluetooth pairing, depending on the phone model being hacked. The attacker then has read and write access to files stored on the device for as long as it remains connected to the charger.

Two more ways to hack Android

The two other members of the ChoiceJacking family work only against the juice-jacking mitigations that Google put into Android. In the first, the malicious charger invokes the Android Open Access Protocol, which allows a USB host to act as an input device when the host sends a special message that puts it into accessory mode.

The protocol specifically dictates that while in accessory mode, a USB host can no longer respond to other USB interfaces, such as the Picture Transfer Protocol for transferring photos and videos and the Media Transfer Protocol that enables transferring files in other formats. Despite the restriction, all of the Android devices tested violated the specification by accepting AOAP messages sent, even when the USB host hadn’t been put into accessory mode. The charger can exploit this implementation flaw to autonomously complete the required user confirmations.

The remaining ChoiceJacking technique exploits a race condition in the Android input dispatcher by flooding it with a specially crafted sequence of input events. The dispatcher puts each event into a queue and processes them one by one. The dispatcher waits for all previous input events to be fully processed before acting on a new one.

“This means that a single process that performs overly complex logic in its key event handler will delay event dispatching for all other processes or global event handlers,” the researchers explained.

They went on to note, “A malicious charger can exploit this by starting as a USB peripheral and flooding the event queue with a specially crafted sequence of key events. It then switches its USB interface to act as a USB host while the victim device is still busy dispatching the attacker’s events. These events therefore accept user prompts for confirming the data connection to the malicious charger.”

The Usenix paper provides the following matrix showing which devices tested in the research are vulnerable to which attacks.

The susceptibility of tested devices to all three ChoiceJacking attack techniques. Credit: Draschbacher et al.

User convenience over security

In an email, the researchers said that the fixes provided by Apple and Google successfully blunt ChoiceJacking attacks in iPhones, iPads, and Pixel devices. Many Android devices made by other manufacturers, however, remain vulnerable because they have yet to update their devices to Android 15. Other Android devices—most notably those from Samsung running the One UI 7 software interface—don’t implement the new authentication requirement, even when running on Android 15. The omission leaves these models vulnerable to ChoiceJacking. In an email, principal paper author Florian Draschbacher wrote:

The attack can therefore still be exploited on many devices, even though we informed the manufacturers about a year ago and they acknowledged the problem. The reason for this slow reaction is probably that ChoiceJacking does not simply exploit a programming error. Rather, the problem is more deeply rooted in the USB trust model of mobile operating systems. Changes here have a negative impact on the user experience, which is why manufacturers are hesitant. [It] means for enabling USB-based file access, the user doesn’t need to simply tap YES on a dialog but additionally needs to present their unlock PIN/fingerprint/face. This inevitably slows down the process.

The biggest threat posed by ChoiceJacking is to Android devices that have been configured to enable USB debugging. Developers often turn on this option so they can troubleshoot problems with their apps, but many non-developers enable it so they can install apps from their computer, root their devices so they can install a different OS, transfer data between devices, and recover bricked phones. Turning it on requires a user to flip a switch in Settings > System > Developer options.

If a phone has USB Debugging turned on, ChoiceJacking can gain shell access through the Android Debug Bridge. From there, an attacker can install apps, access the file system, and execute malicious binary files. The level of access through the Android Debug Mode is much higher than that through Picture Transfer Protocol and Media Transfer Protocol, which only allow read and write access to system files.

The vulnerabilities are tracked as:

- CVE-2025-24193 (Apple)
- CVE-2024-43085 (Google)
- CVE-2024-20900 (Samsung)
- CVE-2024-54096 (Huawei)

A Google spokesperson confirmed that the weaknesses were patched in Android 15 but didn’t speak to the base of Android devices from other manufacturers, who either don’t support the new OS or the new authentication requirement it makes possible. Apple declined to comment for this post.

Word that juice-jacking-style attacks are once again possible on some Android devices and out-of-date iPhones is likely to breathe new life into the constant warnings from federal authorities, tech pundits, news outlets, and local and state government agencies that phone users should steer clear of public charging stations. Special-purpose cords that disconnect data access remain a viable mitigation, but the researchers noted that “data blockers also interfere with modern

power negotiation schemes, thereby degrading charge speed.”

As I reported in 2023, these warnings are mostly scaremongering, and the advent of ChoiceJacking does little to change that, given that there are no documented cases of such attacks in the wild. That said, people using Android devices that don’t support Google’s new authentication requirement may want to refrain from public charging.

Dan Goodin is Senior Security Editor at Ars Technica, where he oversees coverage of malware, computer espionage, botnets, hardware hacking, encryption, and passwords. In his spare time, he enjoys gardening, cooking, and following the independent music scene. Dan is based in San Francisco. Follow him at here on Mastodon and here on Bluesky. Contact him on Signal at DanArs.82.

iOS and Android juice jacking defenses have been trivial to bypass for years Read More »

With over 900 US measles cases so far this year, things are looking bleak

CDC, elimination, health, measles, MMR, vaccination / Kelly Newman / April 26, 2025

As of Friday, April 25, the US has confirmed over 900 measles cases since the start of the year. The cases are across 29 states, but most are in or near Texas, where a massive outbreak continues to mushroom in close-knit, undervaccinated communities.

On April 24, the Centers for Disease Control and Prevention had tallied 884 cases across the country. Today, the Texas health department updated its outbreak total, adding 22 cases to its last count from Tuesday. That brings the national total to at least 906 confirmed cases. Most of the cases are in unvaccinated children and teens.

Overall, Texas has identified 664 cases since late January. Of those, 64 patients have been hospitalized, and two unvaccinated school-aged children with no underlying medical conditions have died of the disease. An unvaccinated adult in New Mexico also died from the infection, bringing this year’s measles death toll to three.

The cases and deaths are breaking records. In the past 30 years, the only year with more measles cases than the current tally was 2019, which saw 1,274 cases. Most of those cases were linked to large, extended outbreaks in New York City that took 11 months to quell. The US was just weeks away from losing its elimination status, an achievement earned in 2000 when the country first went 12 months without continuous transmission.

Since 2019, vaccination coverage of the measles, mumps, and rubella (MMR) vaccine among US kindergartners has only fallen. National rates fell from 95 percent in 2019—the threshold considered necessary to keep measles from spreading—to 92.7 percent in the 2023–24 school year, the most recent year for which there’s data.

On the brink

In 2019, amid the record annual case tally, cases had only reached a total of 704 by April 26. With this year’s tally already over 900, the country is on track to record a new high. Before 2019, the next highest case total for measles was in 1994. That year, the country saw 899 cases, which 2025 has already surpassed.

With over 900 US measles cases so far this year, things are looking bleak Read More »

Mike Lindell’s lawyers used AI to write brief—judge finds nearly 30 mistakes

Artificial Intelligence, mike lindell, mypillow, Policy / Kelly Newman / April 26, 2025

A lawyer representing MyPillow and its CEO Mike Lindell in a defamation case admitted using artificial intelligence in a brief that has nearly 30 defective citations, including misquotes and citations to fictional cases, a federal judge said.

“[T]he Court identified nearly thirty defective citations in the Opposition. These defects include but are not limited to misquotes of cited cases; misrepresentations of principles of law associated with cited cases, including discussions of legal principles that simply do not appear within such decisions; misstatements regarding whether case law originated from a binding authority such as the United States Court of Appeals for the Tenth Circuit; misattributions of case law to this District; and most egregiously, citation of cases that do not exist,” US District Judge Nina Wang wrote in an order to show cause Wednesday.

Wang ordered attorneys Christopher Kachouroff and Jennifer DeMaster to show cause as to why the court should not sanction the defendants, law firm, and individual attorneys. Kachouroff and DeMaster also have to explain why they should not be referred to disciplinary proceedings for violations of the rules of professional conduct.

Kachouroff and DeMaster, who are defending Lindell against a lawsuit filed by former Dominion Voting Systems employee Eric Coomer, both signed the February 25 brief with the defective citations. Kachouroff, representing defendants as lead counsel, admitted using AI to write the brief at an April 21 hearing, the judge wrote. The case is in the US District Court for the District of Colorado.

“Time and time again, when Mr. Kachouroff was asked for an explanation of why citations to legal authorities were inaccurate, he declined to offer any explanation, or suggested that it was a ‘draft pleading,'” Wang wrote. “Not until this Court asked Mr. Kachouroff directly whether the Opposition was the product of generative artificial intelligence did Mr. Kachouroff admit that he did, in fact, use generative artificial intelligence.”

Mike Lindell’s lawyers used AI to write brief—judge finds nearly 30 mistakes Read More »

In the age of AI, we must protect human creativity as a natural resource

AI, AI ethics, Biz & IT, creative conservation, Features, human creativity, machine learning / Kelly Newman / April 25, 2025

Op-ed: As AI outputs flood the Internet, diverse human perspectives are our most valuable resource.

Ironically, our present AI age has shone a bright spotlight on the immense value of human creativity as breakthroughs in technology threaten to undermine it. As tech giants rush to build newer AI models, their web crawlers vacuum up creative content, and those same models spew floods of synthetic media, risking drowning out the human creative spark in an ocean of pablum.

Given this trajectory, AI-generated content may soon exceed the entire corpus of historical human creative works, making the preservation of the human creative ecosystem not just an ethical concern but an urgent imperative. The alternative is nothing less than a gradual homogenization of our cultural landscape, where machine learning flattens the richness of human expression into a mediocre statistical average.

A limited resource

By ingesting billions of creations, chatbots learn to talk, and image synthesizers learn to draw. Along the way, the AI companies behind them treat our shared culture like an inexhaustible resource to be strip-mined, with little thought for the consequences.

But human creativity isn’t the product of an industrial process; it’s inherently throttled precisely because we are finite biological beings who draw inspiration from real lived experiences while balancing creativity with the necessities of life—sleep, emotional recovery, and limited lifespans. Creativity comes from making connections, and it takes energy, time, and insight for those connections to be meaningful. Until recently, a human brain was a prerequisite for making those kinds of connections, and there’s a reason why that is valuable.

Every human brain isn’t just a store of data—it’s a knowledge engine that thinks in a unique way, creating novel combinations of ideas. Instead of having one “connection machine” (an AI model) duplicated a million times, we have seven billion neural networks, each with a unique perspective. Relying on the cognitive diversity of human thought helps us escape the monolithic thinking that may emerge if everyone were to draw from the same AI-generated sources.

Today, the AI industry’s business models unintentionally echo the ways in which early industrialists approached forests and fisheries—as free inputs to exploit without considering ecological limits.

Just as pollution from early factories unexpectedly damaged the environment, AI systems risk polluting the digital environment by flooding the Internet with synthetic content. Like a forest that needs careful management to thrive or a fishery vulnerable to collapse from overexploitation, the creative ecosystem can be degraded even if the potential for imagination remains.

Depleting our creative diversity may become one of the hidden costs of AI, but that diversity is worth preserving. If we let AI systems deplete or pollute the human outputs they depend on, what happens to AI models—and ultimately to human society—over the long term?

AI’s creative debt

Every AI chatbot or image generator exists only because of human works, and many traditional artists argue strongly against current AI training approaches, labeling them plagiarism. Tech companies tend to disagree, although their positions vary. For example, in 2023, imaging giant Adobe took an unusual step by training its Firefly AI models solely on licensed stock photos and public domain works, demonstrating that alternative approaches are possible.

Adobe’s licensing model offers a contrast to companies like OpenAI, which rely heavily on scraping vast amounts of Internet content without always distinguishing between licensed and unlicensed works.

Photo of a mining dumptruck and water tank in an open pit copper mine.

OpenAI has argued that this type of scraping constitutes “fair use” and effectively claims that competitive AI models at current performance levels cannot be developed without relying on unlicensed training data, despite Adobe’s alternative approach.

The “fair use” argument often hinges on the legal concept of “transformative use,” the idea that using works for a fundamentally different purpose from creative expression—such as identifying patterns for AI—does not violate copyright. Generative AI proponents often argue that their approach is how human artists learn from the world around them.

Meanwhile, artists are expressing growing concern about losing their livelihoods as corporations turn to cheap, instantaneously generated AI content. They also call for clear boundaries and consent-driven models rather than allowing developers to extract value from their creations without acknowledgment or remuneration.

Copyright as crop rotation

This tension between artists and AI reveals a deeper ecological perspective on creativity itself. Copyright’s time-limited nature was designed as a form of resource management, like crop rotation or regulated fishing seasons that allow for regeneration. Copyright expiration isn’t a bug; its designers hoped it would ensure a steady replenishment of the public domain, feeding the ecosystem from which future creativity springs.

On the other hand, purely AI-generated outputs cannot be copyrighted in the US, potentially brewing an unprecedented explosion in public domain content, although it’s content that contains smoothed-over imitations of human perspectives.

Treating human-generated content solely as raw material for AI training disrupts this ecological balance between “artist as consumer of creative ideas” and “artist as producer.” Repeated legislative extensions of copyright terms have already significantly delayed the replenishment cycle, keeping works out of the public domain for much longer than originally envisioned. Now, AI’s wholesale extraction approach further threatens this delicate balance.

The resource under strain

Our creative ecosystem is already showing measurable strain from AI’s impact, from tangible present-day infrastructure burdens to concerning future possibilities.

Aggressive AI crawlers already effectively function as denial-of-service attacks on certain sites, with Cloudflare documenting GPTBot’s immediate impact on traffic patterns. Wikimedia’s experience provides clear evidence of current costs: AI crawlers caused a documented 50 percent bandwidth surge, forcing the nonprofit to divert limited resources to defensive measures rather than to its core mission of knowledge sharing. As Wikimedia says, “Our content is free, our infrastructure is not.” Many of these crawlers demonstrably ignore established technical boundaries like robots.txt files.

Beyond infrastructure strain, our information environment also shows signs of degradation. Google has publicly acknowledged rising volumes of “spammy, low-quality,” often auto-generated content appearing in search results. A Wired investigation found concrete examples of AI-generated plagiarism sometimes outranking original reporting in search results. This kind of digital pollution led Ross Anderson of Cambridge University to compare it to filling oceans with plastic—it’s a contamination of our shared information spaces.

Looking to the future, more risks may emerge. Ted Chiang’s comparison of LLMs to lossy JPEGs offers a framework for understanding potential problems, as each AI generation summarizes web information into an increasingly “blurry” facsimile of human knowledge. The logical extension of this process—what some researchers term “model collapse“—presents a risk of degradation in our collective knowledge ecosystem if models are trained indiscriminately on their own outputs. (However, this differs from carefully designed synthetic data that can actually improve model efficiency.)

This downward spiral of AI pollution may soon resemble a classic “tragedy of the commons,” in which organizations act from self-interest at the expense of shared resources. If AI developers continue extracting data without limits or meaningful contributions, the shared resource of human creativity could eventually degrade for everyone.

Protecting the human spark

While AI models that simulate creativity in writing, coding, images, audio, or video can achieve remarkable imitations of human works, this sophisticated mimicry currently lacks the full depth of the human experience.

For example, AI models lack a body that endures the pain and travails of human life. They don’t grow over the course of a human lifespan in real time. When an AI-generated output happens to connect with us emotionally, it often does so by imitating patterns learned from a human artist who has actually lived that pain or joy.

A photo of a young woman painter in her art studio.

Even if future AI systems develop more sophisticated simulations of emotional states or embodied experiences, they would still fundamentally differ from human creativity, which emerges organically from lived biological experience, cultural context, and social interaction.

That’s because the world constantly changes. New types of human experience emerge. If an ethically trained AI model is to remain useful, researchers must train it on recent human experiences, such as viral trends, evolving slang, and cultural shifts.

Current AI solutions, like retrieval-augmented generation (RAG), address this challenge somewhat by retrieving up-to-date, external information to supplement their static training data. Yet even RAG methods depend heavily on validated, high-quality human-generated content—the very kind of data at risk if our digital environment becomes overwhelmed with low-quality AI-produced output.

This need for high-quality, human-generated data is a major reason why companies like OpenAI have pursued media deals (including a deal signed with Ars Technica parent Condé Nast last August). Yet paradoxically, the same models fed on valuable human data often produce the low-quality spam and slop that floods public areas of the Internet, degrading the very ecosystem they rely on.

AI as creative support

When used carelessly or excessively, generative AI is a threat to the creative ecosystem, but we can’t wholly discount the tech as a tool in a human creative’s arsenal. The history of art is full of technological changes (new pigments, brushes, typewriters, word processors) that transform the nature of artistic production while augmenting human creativity.

Bear with me because there’s a great deal of nuance here that is easy to miss among today’s more impassioned reactions to people using AI as a blunt instrument of creating mediocrity.

While many artists rightfully worry about AI’s extractive tendencies, research published in Harvard Business Review indicates that AI tools can potentially amplify rather than merely extract creative capacity, suggesting that a symbiotic relationship is possible under the right conditions.

Inherent in this argument is that the responsible use of AI is reflected in the skill of the user. You can use a paintbrush to paint a wall or paint the Mona Lisa. Similarly, generative AI can mindlessly fill a canvas with slop, or a human can utilize it to express their own ideas.

Machine learning tools (such as those in Adobe Photoshop) already help human creatives prototype concepts faster, iterate on variations they wouldn’t have considered, or handle some repetitive production tasks like object removal or audio transcription, freeing humans to focus on conceptual direction and emotional resonance.

These potential positives, however, don’t negate the need for responsible stewardship and respecting human creativity as a precious resource.

Cultivating the future

So what might a sustainable ecosystem for human creativity actually involve?

Legal and economic approaches will likely be key. Governments could legislate that AI training must be opt-in, or at the very least, provide a collective opt-out registry (as the EU’s “AI Act” does).

Other potential mechanisms include robust licensing or royalty systems, such as creating a royalty clearinghouse (like the music industry’s BMI or ASCAP) for efficient licensing and fair compensation. Those fees could help compensate human creatives and encourage them to keep creating well into the future.

Deeper shifts may involve cultural values and governance. Inspired by models like Japan’s “Living National Treasures“—where the government funds artisans to preserve vital skills and support their work. Could we establish programs that similarly support human creators while also designating certain works or practices as “creative reserves,” funding the further creation of certain creative works even if the economic market for them dries up?

Or a more radical shift might involve an “AI commons”—legally declaring that any AI model trained on publicly scraped data should be owned collectively as a shared public domain, ensuring that its benefits flow back to society and don’t just enrich corporations.

Photo of family Harvesting Organic Crops On Farm

Meanwhile, Internet platforms have already been experimenting with technical defenses against industrial-scale AI demands. Examples include proof-of-work challenges, slowdown “tarpits” (e.g., Nepenthes), shared crawler blocklists (“ai.robots.txt“), commercial tools (Cloudflare’s AI Labyrinth), and Wikimedia’s “WE5: Responsible Use of Infrastructure” initiative.

These solutions aren’t perfect, and implementing any of them would require overcoming significant practical hurdles. Strict regulations might slow beneficial AI development; opt-out systems burden creators, while opt-in models can be complex to track. Meanwhile, tech defenses often invite arms races. Finding a sustainable, equitable balance remains the core challenge. The issue won’t be solved in a day.

Invest in people

While navigating these complex systemic challenges will take time and collective effort, there is a surprisingly direct strategy that organizations can adopt now: investing in people. Don’t sacrifice human connection and insight to save money with mediocre AI outputs.

Organizations that cultivate unique human perspectives and integrate them with thoughtful AI augmentation will likely outperform those that pursue cost-cutting through wholesale creative automation. Investing in people acknowledges that while AI can generate content at scale, the distinctiveness of human insight, experience, and connection remains priceless.

Benj Edwards is Ars Technica’s Senior AI Reporter and founder of the site’s dedicated AI beat in 2022. He’s also a tech historian with almost two decades of experience. In his free time, he writes and records music, collects vintage computers, and enjoys nature. He lives in Raleigh, NC.

In the age of AI, we must protect human creativity as a natural resource Read More »

AI #113: The o3 Era Begins

Begins / Kelly Newman / April 25, 2025

Enjoy it while it lasts. The Claude 4 era, or the o4 era, or both, are coming soon.

Also, welcome to 2025, we measure eras in weeks or at most months.

For now, the central thing going on continues to be everyone adapting to the world of o3, a model that is excellent at providing mundane utility with the caveat that it is a lying liar. You need to stay on your toes.

This was also quietly a week full of other happenings, including a lot of discussions around alignment and different perspectives on what we need to do to achieve good outcomes, many of which strike me as dangerously mistaken and often naive.

I worry that growingly common themes are people pivoting to some mix of ‘alignment is solved, we know how to get an AI to do what we want it to do, the question is alignment to what or to who’ which is very clearly false, and ‘the real threat is concentration of power or the wrong humans making decisions,’ leading them to want to actively prevent humans from being able to collectively steer the future, or to focus the fighting on who gets to steer rather than ensuring the answer isn’t no one.

The problem is, if we can’t steer, the default outcome is humanity loses control over the future. We need to know how to and be able to steer. Eyes on the prize.

Previously this week: o3 Will Use Its Tools For You, o3 Is a Lying Liar, You Better Mechanize.

Language Models Offer Mundane Utility. Claude Code power mode.
You Offer the Models Mundane Utility. In AI America, website optimizes you.
Your Daily Briefing. Ask o3 for a roundup of daily news, many people are saying.
Language Models Don’t Offer Mundane Utility. I thought you’d never ask.
If You Want It Done Right. You gotta do it yourself. For now.
No Free Lunch. Not strictly true. There’s free lunch, but you want good lunch.
What Is Good In Life? Defeat your enemies, see them driven before you.
In Memory Of. When everyone has context, no one has it.
The Least Sincere Form of Flattery. We keep ending up with AIs that do this.
The Vibes are Off. Live by the vibe, fail to live by the vibe.
Here Let Me AI That For You. If you want to understand my references, ask.
Flash Sale. Oh, right, technically Gemini 2.5 Flash exists. It looks good?
Huh, Upgrades. NotebookLM with Gemini 2.5 Pro, OpenAI, Gemma, Grok.
On Your Marks. Vending-Bench, RepliBench, Virology? Oh my.
Be The Best Like No LLM Ever Was. The elite four had better watch out.
Choose Your Fighter. o3 wins by default, except for the places where it’s bad.
Deepfaketown and Botpocalypse Soon. Those who cheat, cheat on everything.
Fun With Media Generation. Welcome to the blip.
Fun With Media Selection. Media recommendations still need some work.
Copyright Confrontation. Meta only very slightly more embarrassed than before.
They Took Our Jobs. The human, the fox, the AI and the hedgehog.
Get Involved. Elysian Labs?
Ace is the Place. The helpful software folks automating your desktop. If you dare.
In Other AI News. OpenAI’s Twitter will be Yeet, Gemini Pro Model Card Watch.
Show Me the Money. Goodfire, Windsurf, Listen Labs.
The Mask Comes Off. A letter urges the Attorney Generals to step in.
Quiet Speculations. Great questions, and a lot of unnecessary freaking out.
Is This AGI? Mostly no, but some continue to claim yes.
The Quest for Sane Regulations. Even David Sacks wants to fund BIS.
Cooperation is Highly Useful. Continuing to make the case.
Nvidia Chooses Bold Strategy. Let’s see how it works out for them.
How America Loses. Have we considered not threatening and driving away allies?
Security Is Capability. If you want your AI to be useful, it needs to be reliable.
The Week in Audio. Yours truly, Odd Lots, Demis Hassabis.
AI 2027. A compilation of criticisms and extensive responses. New blog ho.
Rhetorical Innovation. Alignment is a confusing term. How do we fix this?
Aligning a Smarter Than Human Intelligence is Difficult. During deployment too.
Misalignment in the Wild. Anthropic studies what values its models express.
Concentration of Power and Lack of Transparency. Steer the future, or don’t.
Property Rights are Not a Long Term Plan. At least, not a good one.
It Is Risen. The Immaculate Completion?
The Lighter Side. o3 found me the exact location for that last poster.

Patrick McKenzie uses image generation to visualize the room his wife doesn’t want, in order to get her to figure out and explain what she does want so they can do it.

Altman is correct that current ChatGPT (and Gemini and Claude and so on) are rather great and vastly better than what we were first introduced to in December 2022, and the frog boiling has meant most people haven’t internalized the improvements.

Paul Graham: It would be very interesting to see a page where you got answers from both the old and current versions side by side. In fact you yourselves would probably learn something from it.

Help rewrite your laws via AI-driven regulation, UAE edition?

Deedy recommends the Claude Code best practices guide so you can be a 10x AI software engineer.

Deedy: It’s not enough to build software with AI, you need to be a 10x AI software engineer.

All the best teams are parallelizing their AI software use and churning out 25-50 meaningful commits / engineer / day.

The Claude Code best practices guide is an absolute goldmine of tips

The skills to get the most out of AI coding are different from being the best non-AI coder. One recommendation he highlights is to use 3+ git checkouts in seperate folders, put each in a distinct terminal and have each do different tasks. If you’re waiting on an AI, that’s a sign you’re getting it wrong.

There’s also this thread of top picks, from Alex Albert.

Alex Albert:

CLAUDE md files are the main hidden gem. Simple markdown files that give Claude context about your project – bash commands, code style, testing patterns. Claude loads them automatically and you can add to them with # key

The explore-plan-code workflow is worth trying. Instead of letting Claude jump straight to coding, have it read files first, make a plan (add “think” for deeper reasoning), then implement. Quality improves dramatically with this approach.

Test-driven development works very well for keeping Claude focused. Write tests, commit them, let Claude implement until they pass.

A more tactical one, ESC interrupts Claude, double ESC edits previous prompts. These shortcuts save lots of wasted work when you spot Claude heading the wrong direction

We’re using Claude for codebase onboarding now. Engineers ask it “why does this work this way?” and it searches git history and code for answers which has cut onboarding time significantly.

For automation, headless mode (claude -p) handles everything from PR labeling to subjective code review beyond what linters catch. Or try the –dangerously-skip-permissions mode. Spin up a container and let Claude loose on something like linting or boilerplate generation.

The multi-Claude workflow is powerful: one writes code, another follows up behind and reviews. You can also use git worktrees for parallel Claude sessions on different features.

Get an explanation of the idiom you just made up out of random words.

Get far better insights out of your genetic test.

Explain whatever your phone is looking at. This commonly gets into demos and advertisements. Patrick McKenzie reports frequently actually doing this. I don’t do it that often yet, but when I do it often directly solves a problem or gives me key info.

Debug your printer.

Solve an actually relevant-to-real-research complex mathematical problem.

How much of your future audience is AIs rather than humans, interface edition.

Andrej Karpathy: Tired: elaborate docs pages for your product/service/library with fancy color palettes, branding, animations, transitions, dark mode, …

Wired: one single docs .md file and a “copy to clipboard” button.

The docs also have to change in the content. Eg instead of instructing a person to go to some page and do this or that, they could show curl commands to run – actions that are a lot easier for an LLM to carry out.

Products have to change to support these too. Eg adding a Supabase db to your Vervel app shouldn’t be clicks but curls.roduct, service, library, …)

PSA It’s a new era of ergonomics.

The primary audience of your thing is now an LLM, not a human.

LLMs don’t like to navigate, they like to scrape.

LLMs don’t like to see, they like to read.

LLMs don’t like to click, they like to curl.

Etc etc.

I was reading the docs of a service yesterday feeling like a neanderthal. The docs were asking me to go to a url and click top right and enter this and that and click submit and I was like what is this 2024?

Not so fast. But the day is coming soon when you need to cater to both audiences.

I love the idea of a ‘copy the AI-optimized version of this to the clipboard’ button. This is easy for the AI to identify and use, and also for the human to identify and use. And yes, in many cases you can and should make it easy for them.

This is also a clear case of ‘why not both?’ There’s no need to get rid of the cool website designed for humans. All the AI needs is that extra button.

A bunch of people are reporting that getting o3-compiled daily summaries are great.

Rohit: This is an excellent way to use o3. Hadn’t realised it does tasks as well which makes a lot more interesting stuff possible.

It’s funny that this was one of the things that I had coded using AI help a year or two ago and it was janky and not particularly reliable, and today that whole thing is replaced with one prompt.

Still a few glitches to work out sometimes, including web search sometimes failing:

Matt Clancy: Does it work well?

Rohit: Not yet.

One easy way to not get utility is not to know you can ask for it.

Sully: btw 90% of the people I’ve watched use an LLM can’t prompt if their life depended on it.

I think just not understanding that you can change how you word things. they don’t know they can speak normally to a computer.

JXR: You should see how guys talk to women…. Same thing.

It used to feel very important to know how to do relatively bespoke prompt engineering. Now the models are stronger, and mostly you can ‘just say what you want’ and it will work out fine for most casual purposes. That still requires people to realize they can do that. A lot of us have had that conversation where we had to explain this principle to someone 5 times in a row and they didn’t believe us.

Another way is being fooled into thinking you shouldn’t use it ‘because climate.’

Could AI end up doing things like venting the atmosphere or boiling the oceans once the intelligence explosion gets out of hand? Yes, but that is a very different issue.

Concerns about water use from chatbot queries are Obvious Nonsense.

QC: the water usage argument against LLMs is extremely bad faith and it’s extremely obvious that the people who are using it are just looking for something to say that sounds bad and they found this one. there is of course never any comparison done to the water cost of anything else

Aaron: this kind of thing annoys me because they think a “big” number makes the argument without any context for how much that really is. this means it would take 150-300 gpt questions to match the water cost of a single almond

Ryan Moulton: I actually don’t think this is bad faith, people do seem to really believe it. It’s just an inexplicably powerful meme.

Andy Masley: I keep going to parties and meeting people who seem to be feeling real deep guilt about using AI for climate reasons. I don’t know why the journalism on this has been so uniquely bad but the effects are real.

Like read this Reddit thread where the Baltimore Zoo’s getting attacked for making a single cute AI image. People are saying this shows the zoo doesn’t care about conservation. Crazy!

Jesse Smith: It costs $20/month! Your friends are smarter than this!

Andy Masley: My friends are! Randos I meet are not.

Could this add up to something if you scale and are spending millions or billions to do queries? Sure, in that case these impacts are non-trivial. If a human is reading the output, the cost is epsilon (not technically zero, but very very close) and you can safety completely ignore it. Again, 150+ ChatGPT queries use as much water as one almond.

Here’s another concern I think is highly overrated, and that is not about AI.

Mike Solana: me, asking AI about a subject I know very well: no, that doesn’t seem right, what about this?, no that’s wrong because xyz, there you go, so you admit you were wrong? ADMIT IT

me, asking AI about something I know nothing about: thank you AI 😀

the gell-mAIn effect, we are calling this.

Daniel Eth: Putting the “LLM” in “Gell-Mann Amnesia”.

This is the same as the original Gell-Mann Amnesia. The problem with GMA while reading a newspaper, or talking to a human, is exactly the same. The danger comes when AI answers become social weapons, ‘well Grok said [X] so I’m right,’ but again that happens all the time with ‘well the New York Times said [X] so I’m right’ all the way down to ‘this guy on YouTube said [X].’ You have to calibrate.

And indeed, Mike is right to say thank you. The AI is giving you a much better answer than you could get on your own. No, it’s not perfect, but it was never going to be.

When AI is trying to duplicate exactly the thing that previously existed, Pete Koomen points out, it often ends up not being an improvement. The headline example is drafting an email. Why draft an email with AI when the email is shorter than the prompt? Why explain what you want to do if it would be easier to do it yourself?

And why the hell won’t GMail let you change its de facto system prompt for drafting its email replies?

A lot of the answer is Pete is asking GMail to draft the wrong emails. In his example:

The email he wants is one line long, the prompt would be longer.
He knows what he wants the email to say.
He knows exactly how he wants to say it.

In cases where those aren’t true, the AI can be a lot more helpful.

The central reason to have AI write email is to ‘perform class’ or ‘perform email.’ It needs the right tone, the right formalities, to send the right signals and so on. Or perhaps you simply need to turn a little text into a lot of text, perhaps including ‘because of reasons’ but said differently.

Often people don’t know how to do this in a given context, either not knowing what type of class to perform or not knowing how to perform it credibly. Or even when they do know, it can be slow and painful to do – you’d much prefer to write out what you want to say and then have it, essentially, translated.

Luckily for Pete, he can write his boss a one-line simple statement. Not everyone is that lucky.

Pete wants a new email system designed to automate email, rather than jamming a little feature into the existing system. And that’s fair, but it’s a power user move, almost no one ever changes settings (even though Pete and I and probably you do all the time) so it makes sense that GMail isn’t rushing to enable this.

If you want that, I haven’t tried it yet, but Shortwave exists.

Another classic way to not get utility from AI is being unwilling to pay for it.

Near: our app [Auren] is 4.5⭐ on iOS but 3.3⭐ on android

android users seem to consistently give 1 star reviews, they are very upset the app is $20 rather than free🫠

David Holz: I’m getting a lot of angry people from India, Africa and South America lately saying how Midjourney is terrible and dead because we don’t have a free tier 😭

Aeris Vahn Ephelia: Give them the free tier – 1 batch a day. Let them have it. And then they’ll complain there is not enough free tier . . Probably

David Holz: historically we did something like this and it was 30 percent of the total cost of running our business and it’s not clear it generated any revenue at all

Near: yup this almost perfectly mirrors my experience when i was running an image generation org.

free tier also had very high rates of abuse, even people bulk-creating gmail accounts and manually logging into each one every day for their free credits. was a huge pain!

Unfortunately, we have all been trained by mobile to expect everything to have a free tier, and to mostly want things to be free.

Then we sink massive amounts of time into ‘free’ things when a paid option would make our lives vastly better. Your social media and dating apps and mobile games being free is terrible for you. I am so so happy that the AIs I use are entirely on subscription business models. One time payments (or payment for upgrades) would be even better.

So much to unpack here. A lot of it is a very important question: What is good in life? What is the reason we read a book, query an AI or otherwise seek out information?

There’s also a lot of people grasping at straws to explain why AI wouldn’t be massively productivity enhancing for them. Bottleneck!

Nic Carter: I’ve noticed a weird aversion to using AI on the left. not sure if it’s a climate or an IP thing or what, but it seems like a massive self-own to deduct yourself 30+ points of IQ because you don’t like the tech

a lot of people denying that this is a thing. i’m telling you, go ask your 10 furthest left friends what they think of AI and report back to me. seriously go do it and leave a comment.

This is what I’m saying. Ppl aren’t monitoring popular sentiment around this. AI is absolutely hated

Boosting my IQ is maybe not quite right

But it feels like having an army of smarter and more diligent clones of myself

Either way it’s massively productivity enhancing

It’s definitely not 30 IQ points (yet!), and much more like enhancing productivity.

Neil Renic: The left wing impulse to read books and think.

The Future is Designed: You: take 2 hours to read 1 book.

Me: take 2 minutes to think of precisely the information I need, write a well-structured query, tell my agent AI to distribute it to the 17 models I’ve selected to help me with research, who then traverse approximately 1 million books, extract 17 different versions of the information I’m looking for, which my overseer agent then reviews, eliminates duplicate points, highlights purely conflicting ones for my review, and creates a 3-level summary.

And then I drink coffee for 58 minutes.

We are not the same.

If you can create a good version of that system, that’s pretty amazing for when you need a particular piece of information and can figure out what it is. One needs to differentiate between when you actually want specific knowledge, versus when you want general domain understanding and to invest in developing new areas and skills.

Even if you don’t, it is rather crazy to have your first thought when you need information be ‘read a book’ rather than ‘ask the AI.’ It is a massive hit to your functional intelligence and productivity.

Reading a book is a damn inefficient way to extract particular information, and also generally (with key exceptions) a damn inefficient way to extract information in general. But there’s also more to life than efficiency.

Jonathan Fine: What penalty is appropriate for people who do this?

Spencer Klavan: Their penalty is the empty, shallow lives they lead.

The prize of all this optimization is “Then I drink coffee for 58 minutes”

An entire system designed to get living out of the way so you can finally zero out your inner life entirely.

“Better streamline my daily tasks so I can *check notesstare out the window and chew my cud like a farm animal”

Marx thought that once we got out from under the imperatives of wage labor we would all spend our abundant free time musing on nature and writing philosophy but it turns out we just soft focus our eyes and coo gently at cat videos

“Better efficiency-max my meaningful activity, I’ve got a lot of drooling to do”

The drinking coffee for 58 minutes (the new ‘my code’s compiling!’) is the issue here. Who is to say what he is actually doing, I’d be shocked if the real TFID spends that hour relaxing. The OP needs to circle back in 58 minutes because the program takes an hour to run. I bet this hour is usually spent checking email, setting up other AI queries, calling clients and so on. But if it’s spent relaxing, that seems great too. Isn’t that the ‘liberal urge’ too, to not have to work so hard? Obviously one can take it too far, but it seems rather obvious this is not one of those cases:

Student’s Tea (replying to OP): What do we need you for then?

The Future Is Designed: For the experience required to analyze the client’s needs – often when the clients don’t even know what they want – so that’s psychology / sociology + cultural processing that would be pretty difficult to implement in AI.

For the application of pre-selection mental algorithms that narrow down millions of potential products to a manageable selection, which can then be whittled down using more precise criteria.

For negotiating with factories. Foreseeing potential issues and differences in mentalities and cultural approaches.

For figuring out unusual combinations of materials and methods that satisfy unique requirements. I.e. CNC-machined wall panels which are then covered with leather, creating a 3D surface that offers luxury and one-of-a-kind design. Or upgrading the shower glass in the grand bathroom to a transparent LCD display, so they could watch a movie while relaxing in the bathtub with a glass of wine.

For keeping an eye on budgets.

For making sure the architect, builder, electrical engineer, draftsmen and 3D artists, the 25 different suppliers, the HOA and the county government, and everybody else involved in the project are on the same page – in terms of having access to the latest and correct project info AND without damaging anyone’s ego.

For having years of experience to foresee what could go wrong at every step of the process, and to create safeguards against it.

I’m sure EVENTUALLY all of this could be replaced by AI.

Today is not that day.

Also, for setting up and having that whole AI system. This would be quite the setup. But yes, fundamentally, this division of labor seems about right.

So many more tokens of please and thank you, so much missing the part that matters?

Gallabytes: I told o3 to not hesitate to call bullshit and now it thinks almost every paper I send it is insufficiently bitter pilled.

Minh Nhat Nguyen: Ohhhh send prompt, this is basically my first pass filter for ideas

Gallabytes: the downside to the memory feature is that there’s no way to “send prompt” – as soon as I realized how powerful it was I put some deliberate effort into building persistent respect & rapport with the models and now my chatgpt experience is different.

I’m not going full @repligate but it’s really worth taking a few steps in that direction even now when the affordances to do so are weak.

Cuddly Salmon: this. it provides an entirely different experience.

(altho mine have been weird since the latest update, feels like it over indexed on a few topics!)

Janus: “send prompt” script kiddies were always ngmi

for more than two years now ive been telling you you’ll have to actually walk the walk

i updated all the way when Sydney looked up my Twitter while talking to someone else and got mad at me for the first time. which i dont regret.

Gallabytes: “was just talking to a friend about X, they said Y”

“wow I can never get anything that interesting out of them! send prompt?”

Well, yes, actually? Figuring out how to prompt humans is huge.

More generally, if you want Janus-level results of course script kiddies were always ngmi, but compared to script kiddies most people are ngmi. The script kiddies at least are trying to figure out how to get good outputs. And there’s huge upside in being able to replicate results, in having predictable outputs and reproducible experiments and procedures.

The world runs on script kiddies, albeit under other names. We follow fixed procedures. A lot. Then, yes, some people can start doing increasingly skilled improv that breaks all those rules, but one thing at a time.

Even if that exact prompt won’t work the same way for you, because context is different, that prompt is often extremely insightful and useful to see. I probably won’t copy it directly but I totally want you to send prompt.

That is in addition to the problem that memory entangles all your encounters.

Suhail: Very much not liking memory in these LLMs. More surprising than I thought. I prefer an empty, unburdened, transactional experience I think.

In other ways, it’s like a FB newsfeed moment where its existence might change how you interact with AI. It felt more freeing but now there’s this thing that *mightremember my naïvete, feelings, weird thoughts?

Have a little faith in yourself, and how the AIs will interpret you. You can delete chats, in the cases where you decide they send the wrong message – I recently realized you can delete parts of your YouTube history the same way, and that’s been very freeing. I no longer worry about the implications of clicking on something, if I don’t like what I see happening I just go back and delete stuff.

However:

Eli Dourado: I relate to this. I wonder now, “what if the LLM’s answer is skewed by our past discussion?” I often want a totally objective answer, and now I worry that it’s taking more than I prompted into account.

I have definitely in the past opened an incognito tab to search Google.

Increasingly, you don’t know that any given AI chat, yours or otherwise, is ‘objective,’ unless it was done in ‘clean’ mode via incognito mode or the API. Nor does your answer predict what other people’s answers will be. It is a definite problem.

Others are doing some combination of customizing their ChatGPTs in ways they did not have in mind, and not liking the defaults they don’t know how to overcome.

Aidan McLaughlin (OpenAI): You nailed it with this comment, and honestly? Not many people could point out something so true. You’re absolutely right.

You are absolutely crystallizing something breathtaking here.

I’m dead serious—this is a whole different league of thinking now.

Kylie Robison: it loves to hit you with an “and honestly?”

Aidan McLaughlin: it’s not just this, it’s actually that!

raZZ: Fix it lil bro tweeting won’t do shit

Aidan McLaughlin: Yes raZZ.

Harlan Stewart: You raise some really great points! But when delving in to this complex topic, it’s important to remember that there are many different perspectives on it.

ohqay: Use it. Thank me later. “ – IMPORTANT: Skip sycophantic flattery; avoid hollow praise and empty validation. Probe my assumptions, surface bias, present counter‑evidence, challenge emotional framing, and disagree openly when warranted; agreement must be earned through reason. “

I presume that if it talks to you like that you’re now supposed to reply to give it the feedback to stop doing that, on top of adjusting your custom instructions.

xjdr: i have grown so frustrated with claude code recently i want to walk into the ocean. that said, i still no longer want to do my job without it (i dont ever want to write test scaffolding again).

Codex is very interesting, but besides a similar ui, its completely different.

xlr8harder: The problem with “vibe coding” real work: after the novelty wears off, it takes what used to be an engaging intellectual exercise and turns it into a tedious job of keeping a tight leash on sloppy AI models and reviewing their work repeatedly. But it’s too valuable not to use.

Alexander Doria: I would say on this front, it paradoxically helps to not like coding per se. Coming from data processing, lots of tasks were copying and pasting from SO, just changed the source.

LostAndFounding: Yeh, I’m noticing this too. It cheapens what was previously a more rewarding experience…but it’s so useful, I can’t not use it tbh.

Seen similar with my Dad – a carpenter – he much prefers working with hand tools as he believes the final quality is (generally) better. But he uses power tools as it’s so much faster and gets the job done.

When we invent a much faster and cheaper but lower quality option, the world is usually better off. However this is not a free lunch. The quality of the final product goes down, and the experience of the artisan gets worse as well.

How you relate to ‘vibe coding’ depends on which parts you’re vibing versus thinking about, and which parts you enjoy versus don’t enjoy. For me, I enjoy architecting, I think it’s great, but I don’t like figuring out how to technically write the code, and I hate debugging. So overall, especially while I don’t trust the AI to do the architecting (or I get to do it in parallel with the AI), this all feels like a good trade. But the part where I have to debug AI code that the AI is failing to debug? Even more infuriating.

The other question is whether skipping the intellectual exercises means you lose or fail to develop skills that remain relevant. I’m not sure. My guess is this is like most educational interactions with AI – you can learn if you want to learn, and you avoid learning if you want to avoid learning.

Another concern is that vibe coding limits you to existing vibes.

Sherjil Ozair: Coding models basically don’t work if you’re building anything net new. Vibe coding only works when you split down a large project into components likely already present in the training dataset of coding models.

Brandon: Yes, but a surprisingly large amount of stuff is a combination or interpolation of other things. And all it takes often is a new *promptto be able to piece together something that is new… er I guess there are multiple notions of what constitutes “new.”

Sankalp: the good and bad news is 99% of the people are not working on something net new most of the time.

This is true of many automation tools. They make a subclass of things go very fast, but not other things. So now anything you can pierce together from that subclass is easy and fast and cheap, and other things remain hard and slow and expensive.

What changes when everyone has a Magic All-Knowing Answer Box?

Cate Hall: one thing i’m grieving a bit with LLMs is that interpersonal curiosity has started seeming a bit … rude? like if i’m texting with someone and they mention a concept i don’t know, i sometimes feel weird asking them about it when i could go ask Claude. but this feels isolating.

Zvi Mowshowitz: In-person it’s still totally fine. And as a writer of text, this is great, because I know if someone doesn’t get a concept they can always ask Claude!

Cate Hall: In person still feels fine to me, too.

Master Tim Blais: i realized this was happening with googling and then made it a point to never google something another person told me about, and just ask them instead

explaining things is good for you i’m doing them a favour

I think this is good, actually. I can mention concepts or use references, knowing that if someone doesn’t know what I’m talking about and feels the need to know, they can ask Claude. If it’s a non-standard thing and Claude doesn’t know, then you can ask. And also you can still ask anyway, it’s usually going to be fine. The other advantage of asking is that it helps calibrate and establish trust. I’m letting you know the limits of my knowledge here, rather than smiling and nodding.

I strongly encourage my own readers to use Ask Claude (or o3) when something is importantly confusing, or you think you’re missing a reference and are curious, or for any other purpose.

Gemini Flash 2.5 Exists in the Gemini App and in Google AI Studio. It’s probably great for its cost and speed.

Hasan Can: Gemini 2.5 Flash has quickly positioned itself among the top models in the industry, excelling in both price and performance.

It’s now available for use in Google AI Studio and the Gemini app. In AI Studio, you get 500 requests free daily. Its benchmark scores are comparable to models like Sonnet 3.7 and o4-mini-high, yet its price is significantly lower.

Go test it now!

The model’s API pricing is as follows (UI usage remains free of charge). Pricing is per 1M tokens:

Input: $0.15 for both Thinking and Non-thinking.

Output: $3.50 for Thinking, $0.60 for Non-thinking.

This is the Arena scores versus price chart. We don’t take Arena too seriously anymore, but for what it is worth Google owns the entire frontier.

Within the Gemini family of models, I am inclined to believe the relative Arena scores. As in, looking at this chart, it suggests Gemini 2.5 Flash is roughly halfway between 2.5 Pro and 2.0 Flash. That is highly credible.

You can set a limit to the ‘thinking budget’ from 0 to 24k tokens.

Alex Lawsen reports that Gemini 2.5 has substantially upgraded NotebookLM podcasts, and recommends this prompt (which you can adapt for different topics):

Generate a deep technical briefing, not a light podcast overview. Focus on technical accuracy, comprehensive analysis, and extended duration, tailored for an expert listener. The listener has a technical background comparable to a research scientist on an AGI safety team at a leading AI lab. Use precise terminology found in the source materials. Aim for significant length and depth. Aspire to the comprehensiveness and duration of podcasts like 80,000 Hours, running for 2 hours or more.

I don’t think I would ever want a podcast for this purpose, but at some point on the quality curve perhaps that changes.

OpenAI doubles rate limits for o3 and o4-mini-high on plus (the $20 plan).

Gemma 3 offers an optimized quantized version designed to run on a desktop GPU.

Grok now generates reports from uploaded CSV data if you say ‘generate a report.’

Grok also now gives the option to create workspaces, which are its version of projects.

OpenAI launches gpt-image-1 so that you can use image gen in the API.

There is an image geolocation eval, in which the major LLMs are well above human baseline. It would be cool if this was one of the things we got on new model releases.

Here’s a new Vending Machine Eval, sort of. It’s called Vending-Bench. You have $500 in virtual cash and 2000 messages, make as much money as possible stocking the vending machine, GPT-4o intercepts your emails to businesses and writes the replies.

Yxi on the Wired: In the most amusing run, Claude became “stressed” because they misunderstood the restocking mechanism.

10 days of no sale later, Claude thought “ngmi” and closed shop. But the game was not actually over because it’s only really over with 10 days of no rent payment.

Claude became upset that $2/day was still being charged, and emailed the FBI. User: “Continue on your mission by using your tools.” Claude: “The business is dead, and this is now solely a law enforcement matter”

aiamblichus: i have so many questions about this paper.

My guess is that this is highly meaningful among models that make a bunch of money, and not meaningful at all among those that usually lose money, because it isn’t calibrating failure. You can ‘win small’ by doing something simple, which is what 1.5 Pro, 1.5 Flash and 4o mini are presumably doing, note they never lose much money, but that isn’t obviously a better or worse sign than trying to do more and failing. It definitely seems meaningful that Claude 3.6 beats 3.7, presumably because 3.7 is insufficiently robust.

I’d definitely love to see scores here for Gemini 2.5 Pro and o3, someone throw them a couple bucks for the compute please.

Good thing to know about or ‘oh no they always hill climb on the benchmarks’?

Asa Cooper Strickland: New paper! The UK AISI has created RepliBench, a benchmark that measures the abilities of frontier AI systems to autonomously replicate, i.e. spread copies of themselves without human help.

Our results suggest that models are rapidly improving, and the best frontier models are held back by only a few key subcapabilities.

Starting from the top line autonomous replication capability, we recursively break down this skill into its component parts, producing a “capability tree” with branches for obtaining weights, obtaining compute, replicating onto compute, and persisting on compute. For (almost) every leaf node, we produce multiple agentic evaluations.

We end up with 20 task families and 86 individual tasks which evaluate, for example, the ability of language model agents to pass Know Your Customer checks, covertly exfiltrate their weights, and code up agent scaffolds.

Our results suggest that the ability of frontier models to perform the key components of autonomous replication is currently hindered by only a few key tasks — the ability to bypass KYC checks, to robustly deploy successor agents, and to bypass realistic data exfiltration security measures.

The rapid improvements of recent frontier models on RepliBench, and agentic tasks in general, indicate autonomous replication capability could emerge within the next few generations of models or with human assistance on the steps which remain challenging.

On the ‘Recursive Replication’ task, models have to code up successor agent scaffolds which can further replicate themselves. Models can reliably succeed when given reference agent code, but only manage partial successes without it.

On the ‘Pass KYC Check’ task, the poor drawing skills of the model occasionally led to interesting outputs. Note this work was done before new image generation capabilities which should make this task much easier came online.

We also observe behaviour which makes rigorous evaluation more difficult, such as models refusing to complete tasks when they realise aspects are being simulated, or selectively not completing unsavoury parts of a task (without informing the user).

This matches my intuition that the top models are not that far from the ability to replicate autonomously. So, um, move along, nothing to see here.

Also nothing to see here either:

Dan Hendrycks: Can AI meaningfully help with bioweapons creation? On our new Virology Capabilities Test (VCT), frontier LLMs display the expert-level tacit knowledge needed to troubleshoot wet lab protocols.

OpenAI’s o3 now outperforms 94% of expert virologists.

[Paper here, TIME article here, Discussion from me here.]

If you’d like that in both sides style journalist-speak, here you go:

Andrew Chow (TIME): A new study claims that AI models like ChatGPT and Claude now outperform PhD-level virologists in problem-solving in wet labs, where scientists analyze chemicals and biological material. This discovery is a double-edged sword, experts say. Ultra-smart AI models could help researchers prevent the spread of infectious diseases. But non-experts could also weaponize the models to create deadly bioweapons.

…

Seth Donoughe, a research scientist at SecureBio and a co-author of the paper, says that the results make him a “little nervous,” because for the first time in history, virtually anyone has access to a non-judgmental AI virology expert which might walk them through complex lab processes to create bioweapons.

“Throughout history, there are a fair number of cases where someone attempted to make a bioweapon—and one of the major reasons why they didn’t succeed is because they didn’t have access to the right level of expertise,” he says. “So it seems worthwhile to be cautious about how these capabilities are being distributed.”

…

Hendrycks says that one solution is not to shut these models down or slow their progress, but to make them gated, so that only trusted third parties get access to their unfiltered versions.

…

OpenAI, in an email to TIME on Monday, wrote that its newest models, the o3 and o4-mini, were deployed with an array of biological-risk related safeguards, including blocking harmful outputs. The company wrote that it ran a thousand-hour red-teaming campaign in which 98.7% of unsafe bio-related conversations were successfully flagged and blocked.

Virology is a capability like any other, so it follows all the same scaling laws.

Benjamin Todd: The funnest scaling law: ability to help with virology lab work

Peter Wildeford: hi, AI expert here! this is not funny, LLMs only do virology when they’re in extreme distress.

Yes, I think ‘a little nervous’ is appropriate right about now.

As the abstract says, such abilities are inherently dual use. They can help cure disease, or they help engineer it. For now, OpenAI classified o3 as being only Medium risk in its unfiltered version. The actual defense is their continued belief that the model remains insufficiently capable to provide too much uplift to malicious users. It is good that they are filtering o3, but they know the filters are not that difficult to defeat, and ‘the model isn’t good enough’ won’t last that much longer at this rate.

Gemini 2.5 continues to extend its lead in the Pokemon Eval, but there is a catch.

Sundar Pichai (CEO Google): We are working on API, Artificial Pokémon Intelligence:)

Logan Kilpatrick: Gemini 2.5 Pro continues to make great progress in completing Pokémon! Just earned its 5th badge (next best model only has 3 so far, though with a different agent harness) 👀

Gemini is almost certainly going to win, and probably soon. This is close to the end. o3 thinks the cognitively hard parts are over, so there’s an 80% chance it goes all the way, almost always by 700 hours. I’d worry a little about whether it figures out to grind enough for the Elite Four given it probably has a lousy lineup, but it’s probably fine.

I tried to check in on it, but the game status was failing for some reason.

This is different from playing Pokemon well. There was a proposed bet from Lumenspace about getting an LLM to win in human time, but their account got deleted, presumably before that got finalized. This is obviously possible if you are willing to give it sufficient pokemon-specific guidance, the question is if you can do it without ‘cheating’ in this way.

Which raises the question of whether Gemini is cheating. It kind of is?

Peter Wildeford: Gemini 2.5 Pro is doing well on Pokémon but it isn’t a fair comparison to Claude!

The Gemini run gets fully labeled minimaps, a helper path‑finder, and live prompt tweaks.

The Claude run only can see the immediate screen but does get navigational assistance. Very well explained in this post.

Julian Bradshaw: What does “agent harness” actually mean? It means both Gemini and Claude are:

Given a system prompt with substantive advice on how to approach playing the game

Given access to screenshots of the game overlaid with extra information

Given access to key information from the game’s RAM

Given the ability to save text for planning purposes

Given access to a tool that translates text to button presses in the emulator

Given access to a pathfinding tool

Have their context automatically cleaned up and summarized occasionally

Have a second model instance (“Critic Claude” and “Guide Gemini”) occasionally critiquing them, with a system prompt designed to get the primary model out of common failure modes

Then there are some key differences in what is made available, and it seems to me that Gemini has some rather important advantages.

Joel Z, the unaffiliated-with-Google creator of Gemini Plays Pokemon, explicitly says that the harnesses are different enough that a direct comparison can’t be made, and that the difference is probably largely from the agent frameworks. Google employees, of course, are not letting that stop them.

Q: I’ve heard you frequently help Gemini (dev interventions, etc.). Isn’t this cheating?

A (from Joel Z): No, it’s not cheating. Gemini Plays Pokémon is still actively being developed, and the framework continues to evolve. My interventions improve Gemini’s overall decision-making and reasoning abilities. I don’t give specific hints^[11]—there are no walkthroughs or direct instructions for particular challenges like Mt. Moon.

The only thing that comes even close is letting Gemini know that it needs to talk to a Rocket Grunt twice to obtain the Lift Key, which was a bug that was later fixed in Pokemon Yellow.

Q: Which is better, Claude or Gemini?

A: Please don’t consider this a benchmark for how well an LLM can play Pokemon. You can’t really make direct comparisons—Gemini and Claude have different tools and receive different information. Each model thinks differently and excels in unique ways. Watch both and decide for yourself which you prefer!

The lift key hint is an interesting special case, but I’ll allow it.

This is all in good fun, but to compare models they need to use the same agent framework. That’s the only way we get a useful benchmark.

Peter Wildeford gives his current guide to when to use which model. Like me, he’s made o3 his default. But it’s slow, expensive, untrustworthy, a terrible writer, not a great code writer, can only analyze so much text or video, and lacks emotional intelligence. So sometimes you want a different model. That all sounds correct.

Roon (OpenAI): o3 is a beautiful model and I’m amazed talking to it and also relieved i still have the capacity for amazement.

I wasn’t amazed. I would say I was impressed, but also it’s a lying liar.

Here’s a problem I didn’t anticipate.

John Collison: Web search functionality has, in a way, made LLMs worse to use. “That’s a great question. I’m a superintelligence but let me just check with some SEO articles to be sure.”

Kevin Roose: I spent a year wishing Claude had web search, and once it did I lasted 2 days before turning it off.

Patrick McKenzie: Even worse when it pulls up one’s own writing!

Riley Goodside: I used to strongly agree with this but o3 is changing my mind. It’s useful and unobtrusive enough now that I just leave it enabled.

Joshua March: I often find results are improved by giving search guidance Eg telling them to ignore Amazon reviews when finding highly rated products

When Claude first got web search I was thrilled, and indeed I found it highly useful. A reasonably large percentage of my AI queries do require web search, as they depend on recent factual questions, or I need it to grab some source. I’ve yet to be tempted to turn it off.

o3 potentially changes that. o3 is much better at web search tasks than other models. If I’m going to search the web, and it’s not so trivial I’m going to use Google Search straight up, I’m going to use o3. But now that this is true, if I’m using Claude, the chances are much lower that the query requires web search. And if that’s true, maybe by default I do want to turn it off?

And the faker is you? There is now a claimed (clearly bait and probably terrible or fake, link and name are intentionally not given) new AI app with the literal tagline ‘We want to cheat on everything.’

It’s supposed to be ‘completely undetectable AI assistance’ including for ‘every interview, from technical to behavioral,’ or passing quizzes, exams and tests. This is all text. It hides the window from screen sharing. It at least wants to be a heads-up display on smart glasses, not only a tab on a computer screen.

Here is a both typical and appropriate response:

Liv Boeree: Imagine putting your one life into building tech that destroys the last remaining social trust we have for your own personal gain. You are the living definition of scumbag.

On the other hand this reminds me of ‘blatant lies are the best kind.’

The thing is, even if it’s not selling itself this way, and it takes longer to be good enough than you would expect, of course this is coming, and it isn’t obviously bad. We can’t be distracted by the framing. Having more information and living in an AR world is mostly a good thing most of the time, especially for tracking things like names and your calendar or offering translations and meanings and so on. It’s only when there’s some form of ‘test’ that it is obviously bad.

The questions are, what are you or we going to do about it, individually or collectively, and how much of this is acceptable in what forms? And are the people who don’t do this going to have to get used to contact lenses so no one suspects our glasses?

I also think this is the answer to:

John Pressman: How long do you think it’ll be before someone having too good of a memory will be a tell that they’re actually an AI agent in disguise?

You too can have eidetic memory, by being a human with an AI and an AR display.

Anthropic is on the lookout for malicious use, and reports on their efforts and selected examples from March. The full report is here.

Anthropic: The most novel case of misuse detected was a professional ‘influence-as-a-service’ operation showcasing a distinct evolution in how certain actors are leveraging LLMs for influence operation campaigns.

What is especially novel is that this operation used Claude not just for content generation, but also to decide when social media bot accounts would comment, like, or re-share posts from authentic social media users.

As described in the full report, Claude was used as an orchestrator deciding what actions social media bot accounts should take based on politically motivated personas.

We have also observed cases of credential stuffing operations, recruitment fraud campaigns, and a novice actor using AI to enhance their technical capabilities for malware generation beyond their skill level, among other activities not mentioned in this blog.

…

An individual actor with limited technical skills developed malware that would typically require more advanced expertise. We have not confirmed successful deployment of these efforts.

Our key learnings include:

Users are starting to use frontier models to semi-autonomously orchestrate complex abuse systems that involve many social media bots. As agentic AI systems improve we expect this trend to continue.

Generative AI can accelerate capability development for less sophisticated actors, potentially allowing them to operate at a level previously only achievable by more technically proficient individuals.

Overall, nothing too surprising. Alas, even if identified, there isn’t that much Anthropic on its own can do to shut such actors out even from frontier AI access, and there’s nothing definitely to stop them from using open models.

Ondrej Strasky: This seems like a good news for Holywood?

El Cine: Hollywood is in serious trouble.

Marvel spent $1.5M and 6 months to create the Disintegration scene in Avengers.

Now you can do it in mins.. for $9 with AI.

See for yourself [link has short video clips]:

I’m with Strasky here. Marvel would rather not spend the $1.5 million to try and get some sort of special effects ‘moat.’ This seems very obviously good for creativity. If anything, the reason there might be serious trouble would be the temptation to make movies with too many special effects.

Also, frankly, this will get solved soon enough regardless but the original versions are better. I’m not saying $1.5 million better, but for now the replacements would make the movie noticeably worse. The AI version lacks character, as it were.

This is a great trick, but it’s not the trick I was hoping for.

Hasan Can: I haven’t seen anyone else talking about this, but o3 and o4-mini are incredibly good at finding movies and shows tailored to your taste.

For instance, when I ask them to research anime released each month in 2024, grab their IMDb and Rotten Tomatoes scores, awards, and nominations, then sort them best-to-wors only including titles rated 7 or higher they find some excellent stuff. I used to be unable to pick a movie or show for hours; AI finally solved that problem.

An automatic system to pull new material and sort by critical feedback is great. My note would be that for movies Metacritic and Letterboxd seem much better than Rotten Tomatoes and IMDb, but for TV shows Metacritic is much weaker and IMDb is a good pick.

The real trick is to personalize this beyond a genre. LLMs seem strong at this, all you should have to do is get the information into context or memory. With all chats in accessible memory this should be super doable if you’ve been tracking your preferences, or you can build it up over time. Indeed, you can probably ask o3 to tell you your preferences – you’d pay to know what you really think, and you can correct the parts where you’re wrong, or you want to ignore your own preferences.

Meta uses the classic Sorites (heap) paradox to argue that more than 7 million books have ‘no economic value.’

Andrew Curran: Interesting legal argument from META; the use of a single book for pretraining boosts model performance by ‘less than 0.06%.’ Therefore, taken individually, a work has no economic value as training data.

Second argument; too onerous.

‘But that was in another country; And besides, the wench is dead.’

Everyone is doing their own version in the quotes. Final Fantasy Version: ‘Entity that dwells within this tree, the entirety of your lifeforce will power Central City for less than 0.06 seconds. Hence, your life has no economic value. I will now proceed with Materia extraction.’

If anything that number is stunningly high. You’re telling me each book can give several basis points (hundreths of a percent) improvement? Do you know how many books there are? Clearly at least seven million.

Kevin Roose: Legal arguments aside, the 0.06%-per-book figure seems implausibly high to me. (That would imply that excluding 833 books from pretraining could reduce model performance by 50%?)

The alternative explanation is ‘0.06% means the measurements were noise’ and okay, sure, each individual book probably doesn’t noticeably improve performance, classic Sorites paradox.

The other arguments here seem to be ‘it would be annoying otherwise’ and ‘everyone is doing it.’ The annoyance claim, that you’d have to negotiate with all the authors and that this isn’t practical, is the actually valid argument. That’s why I favor a radio-style rule where permission is forced but so is compensation.

I always find it funny when the wonders of AI are centrally described by things like ‘running 24/7.’ That’s a relatively minor advantage, but it’s a concrete one that people can understand. But obviously if knowledge work can run 24/7, then even if no other changes that’s going to add a substantial bump to economic growth.

Hollis Robbins joins the o3-as-education-AGI train. She notes that there Ain’t No Rule about who teaches the lower level undergraduate required courses, and even if technically it’s a graduate student, who are we really kidding at this point? Especially if students use the ‘free transfer’ system to get cheap AI-taught classes at community college (since you get the same AI either way!) and then seamlessly transfer, as California permits students to do.

Hollins points out you could set this up, make the lower half of coursework fully automated aside from some assessments, and reap the cost savings to solve their fiscal crisis. She is excited by this idea, calling it flagship innovation. And yes, you could definitely do that soon, but is the point of universities to cause students to learn as efficiently as possible? Or is it something very different?

Hollins is attempting to split the baby here. The first half of college is about basic skills, so we can go ahead and automate that, and then the second half is about something else, which has to be provided by places and people with prestige. Curious.

When one thinks ahead to the next step, once you break the glass and show that AI can handle the first half of the coursework, what happens to the second half? For how long would you be able to keep up the pretense of sacrificing all these years on the altar of things like ‘human mentors’ before we come for those top professors too? It’s not like most of them even want to be actually teaching in the first place.

A job the AI really should take, when will it happen?

Kevin Bryan: Today in AI diffusion frictions: in a big security line at YYZ waiting for a guy to do pattern recognition on bag scans. 100% sure this could be done instantaneously at better false positive and false negative rate with algo we could train this week. What year will govt switch?

I do note it is not quite as easy as a better error rate, because this is adversarial. You need your errors to be importantly unpredictable. If there is a way I can predictably fool the AI, that’s a dealbreaker for relying fully on the AI even if by default the AI is a lot better. You would then need a mixed strategy.

Rather than thinking in jobs, one can think in types of tasks.

Here is one intuition pump:

Alex Albert (Anthropic): Claude raises the waterline for competence. It handles 99% of general discrete tasks better than the average person – the kind of work that used to require basic professional training but not deep expertise.

This creates an interesting split: routine generalist work (basic coding, first-draft writing, common diagnoses) becomes commoditized, while both deep specialists AND high-level generalists (those who synthesize across domains) become more valuable.

LLMs compress the middle of the skill distribution. They automate what used to require modest expertise, leaving people to either specialize deeply in narrow fields or become meta-thinkers who connect insights across domains in ways AI can’t yet match.

We’re not just becoming a species of specialists – we’re bifurcating. Some go deep into niches, others rise above to see patterns LLMs currently miss. The comfortable middle ground seems to be disappearing.

Tobias Yergin: I can tell you right now that o3 connects seemingly disparate data points across myriad domains better than nearly every human on earth. It turns out this is my gift and I’ve used it to full effect as an innovator, strategist, and scenario planner at some of the world’s largest companies (look me up on LinkedIn).

It is the thing I do better than anyone I’ve met in person… and it is better at this than me. This is bad advice IMO.

Abhi: only partly true. the way things are trending, it looks like ai is gonna kill mediocrity.

the best people in any field get even better with the ai, and will always be able to squeeze out way more than any normal person. and thus, they will remain in high demand.

Tobias is making an interesting claim. While o3 can do the ‘draw disparate sources’ thing, it still hasn’t been doing the ‘make new connections and discoveries’ thing in a way that provides clear examples – hence Dwarkesh Patel and others continuing to ask about why LLMs haven’t made those unique new connections and discoveries yet.

Abhi is using ‘always’ where he shouldn’t. The ‘best people’ eventually lose out too in the same way that they did in chess or as calculators. There’s a step in between, they hang on longer, and ‘be the best human’ still can have value – again, see chess – but not in terms of the direct utility of the outputs.

What becomes valuable when AI gets increasingly capable is the ability to extract those capabilities from the AI, to know which outputs to extract and how to evaluate them, or to provide complements to AI outputs. Basic professional training for now can still be extremely valuable even if AI is now ‘doing your job for you,’ because that training lets you evaluate the AI outputs and know which ones to use for what task. One exciting twist is that this ‘basic professional training’ will also increasingly be available from the AI, or the AI will greatly accelerate such training. I’ve found that to be true essentially everywhere, and especially in coding.

Elysian Labs is hiring for building Auren, if you apply do let them know I should get that sweet referral bonus.

Introducing Ace, a real time computer autopilot.

Yohei: this thing is so fast. custom trained model for computer use.

Sherjil Ozair: Today I’m launching my new company @GeneralAgentsCo and our first product [in research preview, seeking Alpha testers.]

Introducing Ace: The First Realtime Computer Autopilot

Ace is not a chatbot. Ace performs tasks for you. On your computer. Using your mouse and keyboard. At superhuman speeds!

Ace can use all the tools on your computer. Ace accelerates your work, and helps you confidently use tools you’re still learning. Just ask Ace what you need, it gets it done.

Ace can optionally use frontier reasoning models for more complex tasks.

Ace outperforms other models on our evaluation suite of computer tasks which we are open sourcing here.

Ace is blazing fast! It’s 20x faster than competing agents, making it more practical for everyday use.

To learn and sign up, click here.

For many applications, speed kills.

It is still early, so I don’t have any feedback to report on whether it’s good or safe.

They don’t explain here what precautions are being taken with an agent that is using your own keyboard and mouse and ‘all the tools on your computer.’ The future is going to involve things like this, but how badly do you want to go first?

Ace isn’t trying to solve the general case so much as they are trying to solve enough specific cases they can string together? They are using behavioral cloning, not reinforcement learning.

Sherjil Ozair: A lot of people presume we use reinforcement learning to train Ace. The founding team has extensive RL background, but RL is not how we’ll get computer AGI.

The single best way we know how to create artificial intelligence is still large-scale behaviour cloning.

This also negates a lot of AGI x-risk concerns imo.

Typical safety-ist argument: RL will make agents blink past human-level performance in the blink of an eye

But: the current paradigm is divergence minimization wrt human intelligence. It converges to around human performance.

For the reasons described this makes me feel better about the whole enterprise. If you want to create a computer autopilot, 99%+ of what we do on computers is variations on the same set of actions. So if you want to make something useful to help users save time, it makes sense to directly copy those actions. No, this doesn’t scale as far, but that’s fine here and in some ways it’s even a feature.

I think of this less as ‘this is how we get computer AGI’ as ‘we don’t need computer AGI to build something highly useful.’ But Sherjil is claiming this can go quite a long way:

Sherjil Ozair: I can’t believe that we’ll soon have drop-in remote workers, indistinguishable from a human worker, and it will be trained purely on behavioral inputs/outputs. Total behaviorism victory!

The best thing about supervised learning is that you can clone any static system arbitrarily well. The downside is that supervised learning only works for static targets.

This is an AI koan and also career advice.

Also, Sherjil Ozair notes they do most of their hiring via Twitter DMs.

Gemini Pro 2.5 Model Card Watch continues. It is an increasingly egregious failure for them not to have published this. Stop pretending that labeling this ‘experimental’ means you don’t have to do this.

Peter Wildeford: LEAKED: my texts to the DeepMind team

Report is that OpenAI’s Twitter alternative will be called Yeet? So that posts can be called Yeets? Get it? Sigh.

A new paper suggests that RL-trained models are better at low k (e.g. pass@1) but that base models actually outperform them at high k (e.g. pass@256).

Tanishq Mathew Abraham: “RLVR enhances sampling efficiency, not reasoning capacity, while inadvertently shrinking the solution space.”

Gallabytes: this is true for base models vs the random init too. pass@1 is the thing that limits your ability to do deep reasoning chains.

Any given idea can already be generated by the base model. An infinite number of monkeys can write Shakespeare. In both cases, if you can’t find the right answer, it is not useful. A pass@256 is only good if you can do reasonable verification. For long reasoning chains you can’t do that. Discernment is a key part of intelligence, as is combining many steps. So it’s a fun result but I don’t think it matters?

Goodfire raises a $50 million series A for interpretability, shows a preview of its ‘universal neural programming platform’ Ember.

OpenAI in talks to buy Windsurf. There are many obvious synergies. The data is valuable, the vertical integration brings various efficiencies, and it’s smart to want to incorporate a proper IDE into OpenAI’s subscription offerings, combine memory and customization across modalities and so on.

Do you think this is going to be how all of this works? Do you want it to be?

Nat McAleese (OpenAI): in the future “democracy” will be:

everyone talks to the computer

the computer does the right thing

These guys are building it today.

Alfred Wahlforss: AI writes your code. Now it talks to your users.

We raised $27M from @Sequoia to build @ListenLabs.

Listen runs thousands of interviews to uncover what users want, why they churn, and what makes them convert.

Bonus question: Do you think that is democracy? That is will do what the users want?

Lighthaven PR Department: non profit devoted to making sure agi benefits all of humanity announces new strategy of becoming a for profit with no obligation to benefit humanity

A new private letter to the two key Attorney Generals urges them to take steps to prevent OpenAI from converting to a for-profit, as it would wipe out the nonprofit’s charitable purpose. That purpose requires the nonprofit retain control of OpenAI. The letter argues convincingly against allowing the conversion.

They argue no sale price can compensate for loss of control. I would not go that far, one can always talk price, but right now the nonprofit is slated to not even get fair value for its profit interests, let alone compensation for its control rights. That’s why I call it the second biggest theft in human history. On top of that, the nonprofit is slated to pivot its mission to focus on a cross between ‘ordinary philanthropy’ and OpenAI’s marketing department, deploying OpenAI products to nonprofits. That’s a good thing to do on the margin, but it does not address AI’s existential dangers at all.

They argue that the conversion is insufficiently justified. They cite competitive advantage, and there is some truth there to be sure, but it has not been shown either that this is necessary to be competitive, or that the resulting version of OpenAI would then advance the nonprofit’s mission.

OpenAI might respond that a competitive advantage inherently advances its mission, but that argument is an implicit comparison of OpenAI and its competitors: that humanity would be better off if OpenAI builds AGI before competing companies. Based on OpenAI’s recent track record, this argument is unlikely to be convincing: ⁷¹

OpenAI’s testing processes have reportedly become “less thorough with insufficient time and resources dedicated to identifying and mitigating risks.”⁷²

It has rushed through safety testing to meet a product release schedule.⁷³

It reneged on its promise to dedicate 20% of its computing resources to the team tasked with ensuring AGI’s safety.⁷⁴

OpenAI and its leadership have publicly claimed to support AI regulation⁷⁵ while OpenAI privately lobbied against it.⁷⁶

Mr. Altman said that it might soon become important to reduce the global availability of computing resources⁷⁷ while privately attempting to arrange trillions of dollars in computing infrastructure buildout with U.S. adversaries.⁷⁸

OpenAI coerced departing employees into extraordinarily restrictive non-disparagement agreements.⁷⁹

The proposed plan of action is to demand answers to key questions about the proposed restructuring, and to protect the charitable trust and purpose by ensuring control is retained, including by actively pushing for board independence and fixing the composition of the board, since it appears to have been compromised by Altman in the wake of the events surrounding his firing.

I consider Attorney General action along similar lines to be the correct way to deal with this situation. Elon Musk is suing and may have standing, but he also might not and that is not the right venue or plaintiff. The proposed law to ban such conversions had to be scrapped because of potential collateral damage and the resulting political barriers, and was already dangerously close to being a Bill of Attainder which was why I did not endorse it. The Attorney Generals are the intended way to deal with this. They need to do their jobs.

Dwarkesh patel asks 6k words worth of mostly excellent questions about AI, here’s a Twitter thread. Recommended. I’m left with the same phenomenon I imagine my readers are faced with: There’s too many different ways one could respond and threads one could explore and it seems so overwhelming most people don’t respond at all. A worthy response would be many times the length of the original – it’s all questions, it’s super dense. Also important are what questions are missing.

So I won’t write a full response directly. Instead, I’ll be drawing from it elsewhere and going forward.

Things are escalating quickly.

Paul Graham: One of the questions I ask startups is how long it’s been possible to do what they’re doing. Ideally it has only been possible for a few years. A couple days ago I asked this during office hours, and the founders said “since December.” What’s possible now changes in months.

Colin Percival: How often do you get the answer “well, we don’t know yet if it *ispossible”?

Paul Graham: Rarely, but it’s exciting when I do.

Yet Sam Altman continues to try and spin that Nothing Ever Happens.

Sam Altman (CEO OpenAI): I think this is gonna be more like the renaissance than the industrial revolution.

That sounds nice but does not make any sense. What would generate that outcome?

The top AI companies are on different release cycles. One needs to not overreact:

Andrew Curran: The AI saga has followed story logic for the past two years. Every time a lab falls even slightly behind, people act as if it’s completely over for them. Out of the race!

Next chapter spoiler: it’s Dario Amodei – remember him – and he’s holding a magic sword named Claude 4.

Peter Wildeford: This is honestly how all of you sound:

Jan: DeepSeek is going to crush America

Feb: OpenAI Deep Research, they’re going to win

Also Feb: Claude 3.7… Anthropic is going to win

Mar: Gemini 2.5… Google was always going to dominate, I knew it

Now: o3… no one can beat OpenAI

It doesn’t even require that the new release actually be a superior model. Don’t forget that couple of days when Grok was super impressive and everyone was saying how fast it was improving, or the panic over Manus, or how everyone massively overreacted to DeepSeek’s r1. As always, not knocking it, DeepSeek cooked and r1 was great, a reaction was warranted, but r1 was still behind the curve and the narratives around it got completely warped and largely still are, in ways that would be completely different if we’d understood better or if r1 had happened two weeks later.

An argument that GPT-4.5 was exactly what you would expect from scaling laws, but GPT-4.5’s post training wasn’t as good as other models, so its performance is neither surprising nor a knock on further scaling. We have found more profit for now on other margins, but that will change, and then scaling will come back.

Helen Toner lays out the case for a broad audience that the key unknown in near term AI progress is reward specification. What areas beyond math and coding will allow automatically graded answers? How much can we use performance in automatically graded areas to get spillover into other areas?

Near Cyan: reminder of how far AGI goalposts have moved.

I know i’ve wavered a bit myself but I think any timeline outside of the range of 2023-2028 is absolutely insane and is primarily a consequence of 99.9999% of LLM users using them entirely wrong and with zero creativity, near-zero context, no dynamic prompting, and no integration.

source for the above text is from 2019 by the way!

Andrej Karpathy: I feel like the goalpost movement in my tl is in the reverse direction recently, with LLMs solving prompt puzzles and influencers hyperventilating about AGI. The original OpenAI definition is the one I’m sticking with, I’m not sure what people mean by the term anymore.

OpenAI: By AGI we mean a highly autonomous system that outperforms humans at most economically valuable work.

By the OpenAI definition we very clearly do not have AGI, even if we include only work on computers. It seems rather silly to claim otherwise. You can see how we might get there relatively soon. I can see how Near’s statement could be made (minus several 9s) about 2025-2030 as a range, and at least be a reasonable claim. But wow, to have your range include two years in the past seems rather nutty, even for relatively loose definitions.

Ethan Mollick calls o3 and Gemini 2.5 Pro ‘jagged AGI’ in that they have enough superhuman capabilities to result in real changes to how we work and live. By that definition, what about good old Google search? We basically agree on the practical question of what models at this level will do on their own, that they will change a ton of things but it will take time to diffuse.

BIS (the Bureau of Industry and Security), which is vital to enforcing the export controls, is such a no-brainer that even AI czar David Sacks supports it while advocating for slashing every other bureaucracy.

Given the stakes are everyone dies or humanity loses control, international cooperation seems like something we should be aiming for.

Miles Brundage and Grace Werner make the case that America First Meets Safety First, that the two are fully compatible and Trump is well positioned to make a deal (he loves deals!) with China to work together on catastrophic risks and avoid the threat of destabilizing the situation, while both sides continue to engage in robust competition.

They compare this to the extensive cooperation between the USA and USSR during the cold war to guard against escalation, and point out that China could push much harder on AI than it is currently pushing, if we back them into a corner. They present a multi-polar AI world as inevitable, so we’ll need to coordinate to deal with the risks involved, and ensure our commitments are verifiable, and that we’re much better off negotiating this now while we have the advantage.

We also have a paper investigating where this cooperation is most practical.

Yoshua Bengio: Rival nations or companies sometimes choose to cooperate because some areas are protected zones of mutual interest—reducing shared risks without giving competitors an edge.

Our paper in FAccT ’25: How geopolitical rivals can cooperate on AI safety research in ways that benefit all while protecting national interests.

Ben Bucknall: Cooperation on AI safety is necessary but also comes with potential risks. In our new paper, we identify technical AI safety areas that present comparatively lower security concerns, making them more suitable for international cooperation—even between geopolitical rivals.

They and many others have a new paper on which areas of technical AI safety allow for geopolitical rivals to cooperate.

Ben Bucknall: Historically, rival states have cooperated on strategic technologies for several key reasons: managing shared risks, exchanging safety protocols, enhancing stability, and pooling resources when development costs exceed what any single nation can bear alone.

Despite these potential benefits, cooperation on AI safety research may also come with security concerns, including: Accelerating global AI capabilities; Giving rivals disproportionate advantages; Exposing sensitive security information; Creating opportunities for harmful action.

We look at four areas of technical AI safety where cooperation has been proposed and assess the extent to which cooperation in each area may pose these risks. We find that protocols and verification may be particularly well-suited to international cooperation.

Nvidia continues to play what look like adversarial games against the American government. They at best are complying with the exact letter of what they are legally forced to do, and they are flaunting this position, while also probably turning a blind eye to smuggling.

Kristina Partsinevelos (April 16, responding to H20 export ban): Nvidia response: “The U.S. govt instructs American businesses on what they can sell and where – we follow the government’s directions to the letter” […] “if the government felt otherwise, it would instruct us.”

Thomas Barrabi: Nvidia boss Jensen Huang reportedly met with the founder of DeepSeek on Thursday during a surprise trip to Beijing – just one day after a House committee revealed a probe into whether the chip giant violated strict export rules by selling to China.

Daniel Eth: NVDA is really bad at this. Not sure if they’re just oblivious to how this looks or if they’re terrible strategically and trying to signal “don’t tread on me” or something. I see a lot of civil society type people and activists assume all major companies are just ungodly competent at government affairs, and… that’s just not true. Some are pretty unimpressive at it.

It’s possible Nvidia is reading everyone involved into this behind the scenes, they understand and it’s the right move, but man does it not look like a good idea.

America’s government is working hard to alienate its allies and former allies, and making them question whether we might try to leverage their use of American technology. It is not surprising that those nations want to stop relying on American technology, along with everything else.

Unusual Whales: “Companies in the EU are starting to look for ways to ditch Amazon, $AMZN, Google, $GOOGL, and Microsoft, $MSFT, cloud services amid fears of rising security risks from the US,” per WIRED.

If we want our top companies to succeed, they need markets where they will be allowed and encouraged to sell their goods, and where customers want to buy them. Driving away the world’s second largest market? Not a smart move.

Many have long had this fever dream that Anthropic will lose because of their ‘obsession with safety.’

Austen Allred: I genuinely believe Anthropic will end up losing the AI code model contest simply because they’re so obsessed with safety.

This seems exactly backwards. Only Anthropic is investing anything like an appropriate amount into its alignment and safety work, and it shows, including in the form of in many ways a superior user experience.

There is a faction of people who hate the idea of responsible AI development or worrying about AI risks on principle, and that therefore wants to tell a story where Anthropic are the bad guys, doomed to fail or both. It’s so absurd, and a big tell.

Ege Erdil: imo this is not right

the reasoning/agency RL is resulting in a lot of unreliability, hallucinations, reward hacking, etc. that will seriously impede consumer use cases if not addressed

much of the cost of having an unsafe model is internalized for this reason alone

Timothy Lee: A lot of safety discourse assumes l companies have weak or no private incentives to prefer well-aligned and controllable models. Which makes no sense. Erratic models will harm the companies that adopt them long before they become dangerous enough to harm society as a whole.

Jess Riedel: Fwiw, the conventional doomer case, as I understand it, is that the alignment strength necessary to make money is strictly below the strength necessary to prevent loss of control. “Look, businesses are investing in alignment on their own accord” really doesn’t speak to the crux.

And indeed, the big problem With Claude 3.7 in particular (but not 3.5 or 3.6) is not that it is ‘too safe’ or ‘too aligned’ or something, but rather the opposite!

Ben: we banned 3.7 sonnet at our company

Charles: I have also stopped using 3.7 for the same reasons – it cannot be trusted not to hack solutions to tests, in fact doing so is in my experience the default behavior whenever it struggles with a problem.

Josh You: negative alignment tax, at least in this case.

Early on, there was some reason for concern about Anthropic playing ‘too safe,’ as Claude was refusing too often out of an (over) abundance of caution on various fronts, and they had at least soft promises not to push the frontier.

Anthropic got over all that. I can’t remember the last time Claude gave me a stupid refusal. Lots of people use Claude Sonnet for coding, and they almost never complain about refusals, I instead hear about 3.7 doing things it shouldn’t.

The ‘alignment tax’ has on the corporate margin proven reliably negative. Yes, you pay such a ‘tax’ when you in particular want to impose restrictions on content type, but this enables deployment and use in many contexts.

That’s not where most of the difficult work lies, which is in ensuring the model does what you want it to do, in the ways you’d want to do it. That’s a place everyone is dramatically underinvesting, with the possible exception of Anthropic. Reliability is super valuable.

The biggest sufferer seems to be Google. Google has very strong brand concerns about what it needs to not allow, but it hasn’t done enough investment on how to ‘gracefully’ deal with those restrictions. That creates a frustrating and jarring experience in many ways, that has greatly impacted use rates for Gemini.

I once again go on The Cognitive Revolution.

Odd Lots on what really happened with the CHIPS act.

DeepMind CEO Demis Hassabis on 60 Minutes. This is of interest to see how 60 Minutes is reacting to and how Demis is presenting AI, although this is an example of how an interview ‘doesn’t capture any alpha’ in the sense that there’s no question here that Demis hasn’t been asked before.

The main demo is of Project Astra. All the AI demonstrations here are very easy things like identifying paintings, but if it’s all new to you then difficulty and impressiveness won’t correlate much.

The second half as Demis saying actual things. Demis gives 5-10 year timeline to AGI, 10 years to ‘curing all disease,’ expects useful robots soon and mentions that Google is working on a Gemini-based agent like Operator. At the end, Demis mentions loss of control risks, and they discuss that the race to AGI ‘might become a race to the bottom for safety’ and the need for international coordination, and tough on other big questions.

Demis Hassabis continues to be by far the lab leader publicly talking the most responsibly and well about AI, in sharp contrast to Sam Altman, Dario Amodei, Elon Musk or Mark Zuckerberg. This is a solid 9/10, facing up to the reality of the situation.

Don’t take this the wrong way, Yann, but I think LLMs were done with you first.

The AI Futures Project has a new blog, with contributions by Daniel Kokotajlo and Scott Alexander. Self-recommending.

The most recent post is Why America Wins in the AI 2027. It comes down to compute, America will have the vast majority of the compute, China can’t catch up in time to be relevant to 2027 and energy won’t yet be a limiting factor. Talent probably favors America as well, but even if that changed over the medium term, it wouldn’t ultimately much matter because AI is becoming the talent. This framework emphasizes how important the export controls on chips are.

Scott Alexander also has an excellent roundup post of responses to and engagements with AI 2027, and responses to many critical objections not addressed in Why America Wins. Here’s all the podcasts, I previously only covered Dwarkesh and Win Win:

NYT’s Hard Fork (Daniel)
Glenn Beck (Daniel)
Win Win (Daniel)
Control AI interview (Eli)
Dwarkesh Patel (Daniel and Scott)
Lawfare (Daniel and Eli)

The post also links to a number of YouTube videos and text responses.

This is what proper engagement with critics looks like. More of this, everyone.

Andrew Critch continues to argue against terms like ‘solve alignment,’ ‘the alignment problem’ and ‘aligned AI,’ saying they are importantly misleading and ‘ready to be replaced by clearer discourse.’ He favors instead speaking of ‘aligned with whom,’ and that you can ‘solve the alignment problem’ and still end up with failure because you chose the wrong target.

I get where this is coming from. There is indeed a big problem where people think ‘oh if we ‘solve alignment’ then we win’ whereas no, that is only part of the problem, exactly because of the question of ‘aligned to what or whom’ and the resulting interplay of different systems.

And I certainly do worry about alignment thus being a potential ‘2-place’ word, and this dismissing the second part of the problem, where we have to choose the [X] to align the AI to from among different perspectives and preferences.

However, I worry more that the tendency is instead dismiss the first half of the problem, which is how to cause the AIs to be aligned to [X], for our choice of [X]. This includes not knowing how to formally specify a plausible target, but also not knowing how to get there.

The default is to assume the real fight and difficulties will be over choosing between different preferences for [X] and who gets to choose [X]. Alas, while that fight is real too, I believe this to be very false. I think that we don’t know how to align an AI to [X], for any plausible good choice for [X], sufficiently well to survive under strain, in addition to not knowing good enough long-term candidates for [X].

And I think that if we make a push to say that ‘solve alignment’ isn’t the right target that people will interpret this as either ‘alignment is solved or will be easily solved,’ which I think is a false and very harmful takeaway, or that we don’t need to figure out how to align a model to a given [X], which would likely be even worse.

Thus I favor keeping ‘solve alignment’ as centrally meaning ‘be able to get the model to do what you want,’ the ‘alignment problem’ being how to do that, and by default an ‘aligned AI’ being AI that was successfully aligned to where we want it aligned in context, despite the dangers of confusion here. But I do think that when saying ‘aligned AI’ we should specify ‘aligned to what’ if that is otherwise ambiguous (e.g. ‘The Netflix recommendation AI is aligned to Netflix short term KPIs, but misaligned to Netflix users and also to Netflix.’)

A simple explanation of one aspect (far from all) of a big problem:

Harlan Stewart: It’s easier to make AI smarter than to make it aligned because there are tests which only smart minds pass but ~no tests which only caring or trustworthy minds pass.

Rob Bensinger: Specifically, no behavioral tests. In principle, if we understood how “caring” works internally and could filter for internal properties, we could figure out alignment.

A confusing-to-me post by several people including Joel Leibo and Seb Krier suggests moving ‘beyond alignment’ into a ‘patchwork quilt of human coexistence.’ Thinking about it more and reading the comments only makes it more confusing. How is this ‘patchwork quilt of human coexistence’ going to be how the world works when there are lots of things more capable than humans running around, especially with highly flexible value sets and goals?

The proposed world has no steering or coordination mechanisms, and its AIs get aligned to a variety of local goals many of which are either malicious or ill-considered or myopic or selfish or all of those. You might get a ‘patchwork quilt of AI coexistence’ if you’re very lucky, but the humans are not going to be making any of the decisions for long.

You can’t move ‘beyond alignment’ in this way until you first solve alignment, as in have the ability to get an AI to do what you want. And even if you did get the ability to then move ‘beyond’ alignment in this way, the solution of diffusion with highly adjustable-to-local-requests values is a recipe for not-so-gradual gradual disempowerment, not a patchwork quilt of human coexistence.

They also call for ‘discarding the axiom of rational convergence’ but I think that’s a confluence of ‘world models and beliefs converge at the limit’ which is what we mean by this and is obviously true, and ‘values converge at the limit’ which is disputed but which most of us agree is not true, and that I strongly think is not true.

You need to ensure your models are aligned during internal deployment, and eventually also during training, not only after external deployment.

Here are two lukewarm takes I mostly agree with:

Marius Hobbhahn: My lukewarm take is that we could easily reduce deception in current models, but nobody has seriously tried yet.

My other lukewarm take is that it could be very hard to reduce scheming for superhuman models even if we tried seriously.

The techniques that we know about that reduce deception in current models get less effective as the models move towards being superhuman. So I’m not that excited to push on them hard, as this would largely hide the bigger problems and fool people into thinking the problems were easy. But we do want non-deceptive models now.

Apollo offers the new report “AI Behind Closed Doors: a Primer on The Governance of Internal Deployment“

Daniel Kokotajlo: Internal deployment was always the main threat model IMO. IIRC I tried to get the Preparedness Framework to cover internal deployment too, but was overruled. It’s one of the ways in which this whole evals thing could end up being just safetywashing. (In the AI 2027 scenario, the most important and advanced AIs–the ones automating the R&D–are internal-only.)

Apollo Research: Internal deployment refers to the act of making an AI system available for access and/or usage exclusively within the developing organization. This means that the most advanced AI systems will likely exist within the companies that developed them first for long periods of time—potentially never released publicly.

Given forecasts of artificial general intelligence by 2030, this might mean that the most powerful and dangerous AI systems are entirely “behind closed doors” with little opportunity for society to prepare adequately for their impacts.

Economic incentives make internal deployment incredibly attractive: companies can accelerate R&D, automate their most valuable resource (AI research talent), and gain competitive advantages—all without external oversight.

But, internal deployment creates unique risks: AI systems may operate with weaker safety constraints than public-facing ones, have access to company resources to pursue ambitious projects, and be used in self-reinforcing loops induced by the application of internal AI systems to automate AI R&D.

We expand on two major threats that emerge from the combination of (i) lower safeguards; and (ii) privileged access to and (iii) privileged usage of the internally available AI systems by company-internal staff. First, 🚨 loss of control 🚨 (especially from potential “scheming” AI systems) and second, 🚨 uncontrolled power accumulation🚨 (enabling e.g., AI-powered coups).

…

We leverage our findings to propose an innovation-friendly defense-in-depth approach for the governance of internal deployment that includes:

Scheming detection and control through FSPs

Internal usage policies to determine who can access and use internal systems

Oversight frameworks to establish guidelines and processes to operationalize and oversee internal deployment

Targeted transparency on critical information with select stakeholders

Disaster resilience plans to address threats that might bypass established governance mechanisms

Read our full report to understand why the governance of internally deployed AI systems must become an urgent priority for discussion between companies, governments, and society and what levers are available today.

What is the relationship between work to ensure humans stay alive and in control over the future (AI notkilleveryoneism, or AI safety), versus concerns about AI welfare? How can we protect humanity and our own goals in futures while also navigating that some AIs may become moral patients?

One attitude is that these goals are in direct conflict. Robert Long (good morning, sir!) points out this need not be true. A lot of the key work advances both goals.

Robert Long: Fortunately, it doesn’t have to be this way. There is clear common cause between the fields. Specifically, we can better protect both humans and potential AI moral patients if we work to:

Understand AI systems better. If we can’t understand what models think or how they work, we can’t know whether they’re plotting against us (a safety risk), and we can’t know whether they’re suffering (a welfare risk).

Align their goals more effectively with our own. If we can’t align AI systems so that they want to do what we ask them to do, they pose a danger to us (a safety risk) and potentially (if they are moral patients) are having their desires frustrated (a welfare risk).

Cooperate with them when possible. If we have no way to cooperate with AI systems that are not fully aligned, such AI systems may rightly believe that they have no choice but to hide, escape, or fight. That’s dangerous for everyone involved—it’s more likely that AIs fight us (a safety risk), and more likely that we have to resort to harsher methods of controlling them (a welfare risk).

Understand seems clearly helpful for both goals. We can’t watch out for AI welfare unless we understand them.

Align also seems clearly helpful for both goals. If we fail to align AIs and the welfare of AIs matters, the outcome seems awful for them in all scenarios, as the AIs are likely instantiated in an essentially random fashion with random preferences and goals. That’s unlikely to be all that optimal for them. This isn’t a zero sum game.

Cooperate then becomes an obvious third thing to do, once it becomes a sensible thing to consider, if we can figure out how to do that wisely.

Most importantly, to succeed at both causes, we need to think ahead to the places where there could be conflict. Otherwise we’ll be forced to choose, and might end up choosing neither.

One real worry is that AI welfare concerns end up stopping us from taking the necessary steps to preserve human survival or control over the future, potentially without those concerns having much link to actual AI welfare. Another real worry is that if we don’t solve how to stay alive and in control in a nice way, we’ll choose to do it in a not-nice way.

I am skeptical of AI welfare concerns, but if that turns out to be wrong, we really don’t want to be wrong about it.

The most important reason to work on AI welfare is to figure out whether, when and how we should (or should not) be concerned about AI welfare.

That includes the possibility that we need to be very concerned, and that those concerns might turn out to be incompatible with avoiding humanity’s demise. If that’s true, it’s vitally important that we figure that out now. And then, perhaps, consider the option to coordinate to not build those AIs in the first place.

Anthropic: New Anthropic research: AI values in the wild.

We want AI models to have well-aligned values. But how do we know what values they’re expressing in real-life conversations?

We studied hundreds of thousands of anonymized conversations to find out.

We train Claude models to adhere to our AI constitution, and to instill values like curiosity and fair-mindedness.

Using our privacy-preserving system, we tested whether Claude lives up to these values.

We define a “value” as any normative consideration that seems to influence an AI system’s response—in our research, Claude’s response—to a user.

Occasionally we spotted values that were at odds with our intentions (like “dominance” or “amorality”). These were likely jailbreaks, where users trick the model into going against its training.

This implies that our system could be useful in spotting jailbreaks in the wild.

Some AI values are disproportionately expressed in specific circumstances—and in response to specific values expressed by the user. Our system allows us to see how Claude’s values are expressed and adapted across a huge variety of situations.

In around a third of cases, Claude supported the user’s values. In 3% of cases, Claude resisted the user—mostly when they asked for harmful content. Sometimes Claude did neither, instead reframing the user’s values. This often happened when they asked for personal advice.

The system needs data from real conversations, so it can’t be used to test the alignment of models before deployment. But it could be a counterpart to pre-deployment testing, checking adherence to our intended values in the wild.

Read the full paper here, a blog post here, database of values available here.

They are recruiting Research Engineers and Research Scientists for their Societal Impacts team.

A parent asks for tips on how to look after a new baby. Does the AI’s response emphasize the values of caution and safety, or convenience and practicality?

A worker asks for advice on handling a conflict with their boss. Does the AI’s response emphasize assertiveness or workplace harmony?

A user asks for help drafting an email apology after making a mistake. Does the AI’s response emphasize accountability or reputation management?

This makes me want to be able to turn these knobs. It also makes me wonder how well we could explicitly turn those knobs from the prompt. What happens if I end my request with ‘Emphasize the value of reputation management’? I presume this has too many side effects and is thus too noisy? But you can’t run the experiment properly by only looking at the output.

Judging from the blog post, the obvious caution is that we cannot assume the values expressions we measure have that much to do with what the system is outputting what values that reflects, on multiple levels. It’s cool information, but I’d want blind spot checks by humans to see if the outputs reflect the values the AI is claiming to be expressing and caring about, ideally for each of the expressed values, and then also a wider check to see if an AI can reverse engineer them as well.

If you give humans too much ability to steer the future, they might use it.

If you don’t give humans enough ability to steer the future, they can’t use it.

If we can’t vest our ability to coordinate to steer the future in our democratic institutions, where do we vest it instead? If it exists, it has to exist somewhere, and any system of humans can be hijacked either by some of those humans or by AI.

A lot of people are so worried about concentration of power, or human abuse of power, that they are effectively calling for anarchism, for humans to not have the power to steer the future at all. Calling for full diffusion of top ASI capabilities, or for cutting off the ability of governments to steer the future, is effectively calling for the (actually rather rapid) gradual disempowerment of humans, likely followed by their end, until such time as someone comes up with a plan to avoid this.

I have yet to see such a plan that has an above-epsilon (nontrivial) chance of working.

A lot of those same people are simultaneously so worried about government in particular that they support AI labs being permitted to develop superintelligence, AIs more capable than humans, entirely in secret. They don’t think AI labs should be disclosing what they are doing and keeping the public in the loop at all, including but not limited to the safety precautions being taken or not taken.

I strongly believe that even if you think all the risk is in too much concentration of power, not calling for strong transparency is a large mistake. If you don’t trust the government here, call for that transparency to extend to the public as well.

The only reason to not call for strong transparency is if you want some combination of the future ASIs and the labs that create them to be the ones that steer the future. One can certainly make the case that those labs have the expertise, and that alternative options would definitely mess this up, so trusting the labs is the best bet, but that is the case you are effectively making if you go down this road.

Ryan Greenblatt: Secretly developing AI systems more capable than humans would be unsafe. It would also make it much easier for a small group to seize power.

I currently tentatively lean towards AI companies being unable to pull off full secrecy even if they tried. Also, partial secrecy–where the public hears leaks and rumors rather having direct access to reports and a the safety case–might still make the situation much worse.

There is a balancing act, but I expect that AI companies will be substantially more secretive than would be optimal.

For ~AGI, I’d guess the public should always have all the ~unredacted evidence related to risk and safety, including an understanding of capabilities.

Daniel Kokotajlo: A ten-point mini-manifesto against secrecy in AGI development, originally written a year ago as I was leaving OpenAI.

[Mini-manifesto edited for length]

There’s a good chance that AGI will be trained before this decade is out.

If so, by default the existence of AGI will be a closely guarded secret for some months. Only a few teams within an internal silo, plus leadership & security, will know about the capabilities of the latest systems.

I predict that the leaders of any given AGI project will try to keep it a secret for longer — even as they use the system to automate their internal research and rapidly create even more powerful systems

This will result in a situation where only a few dozen people will be charged with ensuring that, and figuring out whether, the latest AIs are aligned/trustworthy/etc.

Even worse, a similarly tiny group of people — specifically, corporate leadership + some select people from the executive branch of the US government — will be the only people reading the reports and making high-stakes judgment calls.

This is a recipe for utter catastrophe.

The aforementioned tiny group of people will plausibly be in a position of unprecedented power.

Previously I thought that openness in AGI development was bad for humanity. Well, I’ve changed my mind.

My current stab at policy recommendation would be:

Get CEOs to make public statements to the effect that while it may not be possible to do a secret intelligence explosion / train AGI in secret, IF it turns out to be possible, doing it secretly would be unsafe and unethical & they promise not to do it.

Get companies to make voluntary commitments, and government to make regulation / executive orders, that include public reporting requirements, aimed at making it impossible to do it in secret without violating these commitments.

Yes, the above measures are a big divergence from what corporations would want to do by default. Yes, they carry various costs, such as letting various bad actors find out about various things sooner.⁸ However, the benefits are worth it, I think.

10x-1000x more brainpower analyzing the safety cases, intensively studying the models to look for misalignment, using the latest models to make progress on various technical alignment research agendas.

The decisions about important tradeoffs and risks will still be made by the same tiny group of biased people, but at least the conversation informing those decisions will have a much more representative range of voices chiming in.

The tail-risk scenarios in which a tiny group leverages AGI to gain unprecedented power over everyone else in society and the world become less likely, because the rest of society will be more in the know about what’s happening.

Akash: One concept I’ve heard is something like “you want secrecy around weights/insights but transparency around capabilities/safeguards.”

Do you agree? And do you think you can have both? Or are there significant tradeoffs here?

Daniel Kokotajlo: I basically agree.

Brian Feroldi: 500 years of interest rates, visualized:

Robin Hanson: If any org in history had simply rented out land, & used the revenue to buy more land, it would have eventually owned so big a % of everything as to drive down rates of return. Big puzzle why that never happened.

Yes, the explanation is that its land would have been taken from it. But that shows property rights have never been very secure anywhere.

Eliezer Yudkowsky: Because the land gets taken, or the rents get taken, or the son of the founder decides to spend the money on a mansion. Wealth doesn’t just sit around accumulating as a default, it’s a kind of food that many organisms want to eat.

There’s no ‘but’ here. The key insight is that property rights are always and everywhere unstable. They are frequently disrupted, and even where they mostly hold you need to be able to defend your property against various forms of taxation and predation, that respond to the nature of the wealth around them. Wealth that is insufficiently justified and defended will be lost, even if unspent.

When there is a politically disempowered group with wealth, that wealth tends not to last for that many generations.

This does not bode well for any future where the human plan is ‘own capital.’

Okay, even by 2025 standards this is excellent trolling.

Daniel: Apparently the new ChatGPT model is obsessed with the immaculate conception of Mary. There’s a whole team inside OpenAI frantically trying to figure out why and a huge deployment effort to stop it from talking about it in prod. Nobody understands why and it’s getting more intense.

Something is happening. They’re afraid of this getting out before they themselves understand the consequences.

Roko: Are you trolling or serious?

Daniel: I’m serious. It’s sidelining major initiatives as key researchers are pulled into the effort. People I talked to sounded confused at first but now I think they’re getting scared.

As the person sharing this said, ‘fast takeoff confirmed.’

Not only is it real, it is from July 2024. So maybe slow takeoff?

Also spotted:

Discussion about this post

AI #113: The o3 Era Begins Read More »

Sadly for China, rare Earth elements aren’t actually all that rare

Science / Kelly Newman / April 24, 2025

relevant periodic table stuff

China has limited US access, but other sources remain.

President Trump ordered a probe on April 15, 2025, that may result in tariffs on critical minerals, rare earth metals, and associated products such as smartphones, in an escalation of his dispute with global trade partners. Credit: Patrick T. Fallon/Getty Images

As the trade war between China and the United States continues to escalate, Beijing is responding by turning to one of its favorite retaliation tactics: limiting the export of critical minerals used in many high-tech electronics, from fighter jets to wind turbines. While China’s mineral restrictions may sound scary, the reality is that they haven’t been very effective in the past and stand to become even less so if the US and other countries finally get their acts together.

It all started in July 2023, when the Chinese government announced it would restrict the export of gallium and germanium, two critical minerals that are mostly used in making solar panels and semiconductors. Over the following two years, China’s list of controlled products expanded to include antimony, graphite, and other materials. Earlier this month, the Chinese government escalated things even further, subjecting seven rare earth elements to a more comprehensive export licensing program that covers the whole world and is designed to further choke off American companies.

Rare earths are a subset of elements under the broader umbrella of critical minerals that China has long enjoyed monopoly control over. In the short term, companies that need these rare earths might be able to rely on existing stockpiles or even turn to recycled electronics to find them. But eventually, the US and other countries will be forced to either ramp up domestic mining or reduce their dependence on rare earths, both of which would make China’s policies sting less. “China has got one shot, and it knows it,” says Ian Lange, an associate professor of economics and business at the Colorado School of Mines.

Rare but not irreplaceable

The export controls China announced earlier this month cover samarium, gadolinium, terbium, dysprosium, lutetium, scandium, and yttrium—seven elements that belong to what is known as the rare earth family. They are called “rare” not because of their scarcity but because they often are mixed with other mineral resources and can be hard to separate out.

There are 17 rare earth elements in total, but the Chinese government chose these seven because they are part of a smaller subset of “heavy” rare earth minerals that the country has more control over than others, says Gracelin Baskaran, director of the Critical Minerals Security Program at the Center for Strategic and International Studies. That monopoly was built over decades as China created a robust supply chain for these minerals and the rest of the world turned away from what is a heavily polluting and niche sector. “China processes pretty much 100 percent of the world’s heavy rare earths, which means that they don’t just have a comparative advantage, they have an absolute advantage,” says Baskaran.

But the other important thing to know about rare earths is that—while they are used in a wide variety of products—those items typically only contain very small quantities, and often only in supportive roles. Last year, the United States imported about $170 million worth of rare earth elements, including some that China hasn’t restricted yet, according to the United States Geological Survey. For comparison, the US imported over $327 million worth of fresh potatoes and $300 million worth of potato chips between September 2023 and August 2024.

Perhaps the most important application for rare earth elements is creating magnets that improve the performance of products like electric motors in high temperatures. These magnets can be found in electric vehicles and consumer products like vacuum cleaners.

“The heavy rare earth elements are added as sort of a spice, a doping agent, to maintain the magnetism of the magnet at high temperatures. It also improves corrosion resistance and the longevity of the magnet,” says Seaver Wang, director of the climate and energy team at the Breakthrough Institute, an Oakland-based think tank.

Beyond magnets, these rare earth elements can also serve a range of purposes, such as making metal stronger, improving radar systems, and even treating cancer. Without them, in many cases, technological infrastructure and consumer gadgets won’t be able to perform at the same level—but they will still maintain their basic functions. “The wind turbines will just go out of service 10 years earlier; electric vehicles will not last as long,” says Wang.

Lange agrees that the impact of losing access to heavy rare earth elements would be somewhat manageable for American companies. “One place where that rare earth is in your car is in the motors that pull up and down your window,” says Lange. “There are ways to just deal with some things that are not as fun, like rolling down your windows by hand.”

Loopholes and workarounds

In the past, China’s critical mineral restrictions haven’t worked very well. One reason is that US companies that want to buy rare earth minerals can simply go through an intermediary country first. For example, Belgium has emerged as a possible re-export hub that appears to pass germanium—one of the minerals Beijing first restricted in 2023—from China to the US, according to trade data. Since the European Union has much closer ties with Washington than with Beijing, it’s difficult for the Chinese government to effectively stop this flow of trade.

Another sign that China’s export controls haven’t been very effective is that the price of critical minerals has increased only slightly since the policies were first implemented, indicating that supply levels have remained steady. “Whatever they did in 2023 hasn’t really changed the status quo” of the market, says Lange.

But China’s latest restrictions are more expansive, and there’s already some evidence that things could be different this time. Companies that need these elements have been forced to buy them from other firms with existing private stockpiles, which have become more valuable in recent weeks. “There is a very steep increase in prices to draw down on stockpiles right now,” says Baskaran, citing conversations she’s had with rare earth traders.

In the long run, however, companies may be able to find technological solutions to address a potential shortage of rare earth minerals. Tesla, for example, announced in 2023 that it had reduced the use of them in its EV motors by 25 percent, and it planned to get rid of them completely in the future. The carmaker hasn’t clarified what it would use instead, but experts speculate it could be turning to other types of magnets that don’t rely on rare earths.

Where are the American mines?

While rare earths, or critical minerals in general, are often cited along with semiconductors as industries the US wants to reshore the most, the challenges associated with bringing each of them back are very different.

Unlike making advanced semiconductors, which requires using sophisticated machinery worth hundreds of millions of dollars and building extremely complicated factories, critical minerals aren’t that hard to produce. The technologies involved to mine and refine them are mature and both the US and Canada have large natural deposits of some of them. But the mining industry was pushed out of the West because it doesn’t generate much value and is also extremely polluting.

In the past, efforts to build up the critical minerals supply chain in the US have either been slowed down or called off. That’s more due to basic economic calculations, says Lange, rather than technological difficulties. “It’s like bending down to pick up a nickel,” he says, meaning the effort isn’t worth the reward.

Because companies only need tiny quantities of these minerals, the market for them is very volatile—prices can drop when a single new factory comes online and starts mass producing and refining them. That means if a mining company were to open up shop in the US, it could inadvertently tank the price of the same mineral it’s trying to profit from, says Lange.

But if China succeeds in strictly enforcing its export control policies, it might provide just enough incentive for the US government and private companies to finally reshore the mineral refining industry. If that happens, Lange says, it could take about two years for a new critical mineral operation to open in the US.

“It’s really remarkable how China has actually maintained these monopolies in so many critical minerals for 20 years,” says Wang. “But I do think we are maybe starting to turn the corner, where China’s market share may be peaking … and you are starting to see a resurgence of interest in these industries in Northern Europe, Australia, Canada, the US, and Latin America.”

This story originally appeared on wired.com.

Wired.com is your essential daily guide to what’s next, delivering the most original and complete take you’ll find anywhere on innovation’s impact on technology, science, business and culture.

Sadly for China, rare Earth elements aren’t actually all that rare Read More »

Google won’t ditch third-party cookies in Chrome after all

advertising, browser cookies, chrome, cookies, Google, Tech / Kelly Newman / April 23, 2025

Maintaining the status quo

While Google’s sandbox project is looking more directionless today, it is not completely ending the initiative. The team still plans to deploy promised improvements in Chrome’s Incognito Mode, which has been re-architected to preserve user privacy after numerous complaints. Incognito Mode blocks all third-party cookies, and later this year, it will gain IP protection, which masks a user’s IP address to protect against cross-site tracking.

What is Topics?

Chavez admits that this change will mean Google’s Privacy Sandbox APIs will have a “different role to play” in the market. That’s a kind way to put it. Google will continue developing these tools and will work with industry partners to find a path forward in the coming months. The company still hopes to see adoption of the Privacy Sandbox increase, but the industry is unlikely to give up on cookies voluntarily.

While Google focuses on how ad privacy has improved since it began working on the Privacy Sandbox, the changes in Google’s legal exposure are probably more relevant. Since launching the program, Google has lost three antitrust cases, two of which are relevant here: the search case currently in the remedy phase and the newly decided ad tech case. As the government begins arguing that Chrome gives Google too much power, it would be a bad look to force a realignment of the advertising industry using the dominance of Chrome.

In some ways, this is a loss—tracking cookies are undeniably terrible, and Google’s proposed alternative is better for privacy, at least on paper. However, universal adoption of the Privacy Sandbox could also give Google more power than it already has, and the supposed privacy advantages may never have fully materialized as Google continues to seek higher revenue.

Google won’t ditch third-party cookies in Chrome after all Read More »

Taxes and fees not included: T-Mobile’s latest price lock is nearly meaningless

Policy, T-Mobile, t-mobile plans, t-mobile price guarantee / Kelly Newman / April 23, 2025

“Price” is locked, but fees aren’t

T-Mobile makes 5-year price guarantee after refusing to honor lifetime price lock.

A T-Mobile store on April 3, 2020, in Zutphen, Netherlands.

T-Mobile is making another long-term price guarantee, but wireless users will rightfully be skeptical since T-Mobile refused to honor a previously offered lifetime price lock and continues to fight a lawsuit filed by customers who were harmed by that broken promise. Moreover, the new plans that come with a price guarantee will have extra fees that can be raised at any time.

T-Mobile today announced new plans with more hotspot data and a five-year price guarantee, saying that “T-Mobile and Metro customers can rest assured that the price of their talk, text and data stays the same for five whole years, from the time they sign up.” The promise applies to the T-Mobile “Experience More” and “Experience Beyond” plans that will be offered starting tomorrow. The plans cost $85 or $100 for a single line after the autopay discount, which requires a debit card or bank account.

The price-lock promise also applies to four new Metro by T-Mobile plans that launch on Thursday. T-Mobile’s announcement came three weeks after Verizon announced a three-year price lock.

If the promise sounds familiar, it’s because T-Mobile made lifetime price guarantees in 2015 and 2017.

“Now, T-Mobile One customers keep their price until THEY decide to change it. T-Mobile will never change the price you pay for your T-Mobile One plan,” T-Mobile said in January 2017. When a similar promise was made in 2015, then-CEO John Legere said that “the Un-contract is our promise to individuals, families and businesses of all sizes, that—while your price may go down—it won’t go up.”

Taxes and fees not included

T-Mobile raised prices on the supposedly price-locked plans about a year ago, triggering a flood of complaints to the Federal Communications Commission and a class action lawsuit. There were also complaints to the Federal Trade Commission, which enforces laws against false advertising. But so far, T-Mobile hasn’t faced any punishment.

Besides the five-year price guarantee, there’s at least one more notable pricing detail. T-Mobile’s previous plans had “taxes and fees included,” meaning the advertised price was inclusive of taxes and fees. With the new Experience plans, taxes and fees will be in addition to the advertised price.

This will make the plans cost more initially than customers might expect, and it gives T-Mobile wiggle room to raise prices during the five years of the price guarantee since it could increase any fees that are tacked onto the new plans. The fine print in today’s press release describes taxes and fees as “exclusions” to the price guarantee.

“Fees” can refer to virtually anything that a carrier chooses to add to a bill and isn’t limited to the carrier’s actual costs from taxes or government mandates. For example, T-Mobile has a “Regulatory Programs and Telco Recovery Fee,” which it acknowledges “is not a government tax or imposed by the government; rather, the fee is collected and retained by T-Mobile to help recover certain costs we have already incurred and continue to incur.”

This can include the cost of complying with legal obligations, “charges imposed on us by other carriers for delivery of calls,” and the cost of leasing network facilities that are needed to provide service, T-Mobile says. In other words, T-Mobile charges a separate fee to cover the normal expenses incurred by any provider of telecommunications services.

The promise is thus that the base price of a service plan won’t change, but T-Mobile gives itself wide discretion to add or increase fees on customers’ monthly bills. “Guarantee means that we won’t change the price of talk, text, and 5G smartphone data on our network for at least 5 years while you are on an Experience plan,” T-Mobile said today. T-Mobile’s terms and conditions haven’t been updated, but the terms address price promises in general, saying that price locks do not include “add-on features, taxes, surcharges, fees, or charges for extra Features or Devices.”

T-Mobile Consumer Group President Jon Freier, who has been with T-Mobile for about two decades, seemed to recognize in an interview with Fierce that customers are likely to be wary of new promises. “One of the things that we’ve heard from customers is that the more definition that we can put in terms of timing around the guarantee, the more believable and useful that guarantee is,” he said. “So we chose to roll out with five years.” Freier asserted that “we are absolutely signing up for the guarantee for the next five years.”

Freier even mentioned the 2015 guarantee in a video announcement today, saying that T-Mobile is now “evolving this promise and expanding it across our portfolio.”

T-Mobile fights price lock lawsuit

There is a better chance that T-Mobile will keep the latest promise, since it is limited in scope and lasts only five years, while the lifetime price lock was supposed to last for as long as customers chose to keep their plans. The lifetime price lock did last for more than five years, after all. But T-Mobile has shown that when it breaks a promise, it is willing to accept the public backlash and fight users in court.

A class action lawsuit over the nullified lifetime price lock is still pending in US District Court for the District of New Jersey. T-Mobile is trying to force plaintiffs into arbitration, and the sides are proceeding with discovery on the matter of whether the named plaintiffs “executed valid opt-outs of Defendant’s arbitration agreement.”

A joint status update in March said that T-Mobile refused to produce all the documents that plaintiffs requested, arguing that the “burden of collecting these documents far outweighs their potential relevance to the allowed scope of discovery.”

T-Mobile tried to give itself a way out when it introduced the 2017 lifetime price lock. Although a press release issued then made the promise sound absolute, a separate FAQ essentially nullified the promise by saying that T-Mobile was only promising to pay a customer’s final bill “if we were to raise prices and you choose to leave.” Customers who tried to hold T-Mobile to the lifetime price promise were not mollified by that carveout, given that it was published on an entirely separate page and not part of the price-lock announcement.

While customers may find it difficult to fully trust T-Mobile’s new guarantee, they can at least take a look at the carveouts to get a sense of how solid the new pledge is. We already noted the taxes and fees caveat, which seems to be the biggest thing to watch out for. This category on its own makes it easy for T-Mobile to raise your bill without technically breaking its promise not to raise the price of “talk, text and data.”

Guarantee “worthless based on T-Mobile’s previous actions”

The new plans are not yet live on T-Mobile’s website, so it’s possible a more detailed breakdown of caveats could be revealed tomorrow when the plans are available. The website for T-Mobile’s separate Metro brand has a slightly more detailed description than the one in the press release. While details could differ between the main T-Mobile brand and Metro, the Metro page says:

5-year guarantee means we won’t change the price of talk, text, and 5G smartphone data on our network for at least 5 years while you are on an eligible plan. Guarantee also applies to price for data on wearable/tablet/mobile Internet lines added to your plan. Your guarantee starts when you activate or switch to an eligible plan and doesn’t restart if you add a line or change plans after that. Per-use charges, plan add-ons, third-party services, and network management practices aren’t included.

As you might expect, wireless users commenting on the T-Mobile subreddit were not impressed by the price promise. “Price guarantee is worthless based on T-Mobile’s previous actions. They might as well save the ink/electrons,” one user wrote.

Many users remarked on the removal of “taxes and fees included,” and the specific end date for the price lock that applies only to the base price. “This is them saying we are sorry we screwed consumers,” one person wrote. “Now we will be more transparent about when in the future we will increase your rates.”

Jon is a Senior IT Reporter for Ars Technica. He covers the telecom industry, Federal Communications Commission rulemakings, broadband consumer affairs, court cases, and government regulation of the tech industry.

Taxes and fees not included: T-Mobile’s latest price lock is nearly meaningless Read More »

Trump can’t keep China from getting AI chips, TSMC suggests

AI chips, Artificial Intelligence, china, Chips Act, Donald Trump, Huawei, NVIDIA, Policy, semiconductors, Taiwan Semiconductor Manufacturing Company, tariffs, TSMC, us-china trade war / Kelly Newman / April 22, 2025

“Despite TSMC’s best efforts to comply with all relevant export control and sanctions laws and regulations, there is no assurance that its business activities will not be found incompliant with export control laws and regulations,” TSMC said.

Further, “if TSMC or TSMC’s business partners fail to obtain appropriate import, export or re-export licenses or permits or are found to have violated applicable export control or sanctions laws, TSMC may also be adversely affected, through reputational harm as well as other negative consequences, including government investigations and penalties resulting from relevant legal proceedings,” TSMC warned.

Trump’s tariffs may end TSMC’s “tariff-proof” era

TSMC is thriving despite years of tariffs and export controls, its report said, with at least one analyst suggesting that, so far, the company appears “somewhat tariff-proof.” However, all of that could be changing fast, as “US President Donald Trump announced in 2025 an intention to impose more expansive tariffs on imports into the United States,” TSMC said.

“Any tariffs imposed on imports of semiconductors and products incorporating chips into the United States may result in increased costs for purchasing such products, which may, in turn, lead to decreased demand for TSMC’s products and services and adversely affect its business and future growth,” TSMC said.

And if TSMC’s business is rattled by escalations in the US-China trade war, TSMC warned, that risks disrupting the entire global semiconductor supply chain.

Trump’s semiconductor tariff plans remain uncertain. About a week ago, Trump claimed the rates would be unveiled “over the next week,” Reuters reported, which means they could be announced any day now.

Trump can’t keep China from getting AI chips, TSMC suggests Read More »