Author name: Mike M.

Right-wing political violence is more frequent, deadly than left-wing violence

hate speech, right-wing extremism, Science, Social science, syndication / Mike M. / September 18, 2025

President Trump’s assertions about political violence ignore the facts.

After the Sept. 10, 2025, assassination of conservative political activist Charlie Kirk, President Donald Trump claimed that radical leftist groups foment political violence in the US, and “they should be put in jail.”

“The radical left causes tremendous violence,” he said, asserting that “they seem to do it in a bigger way” than groups on the right.

Top presidential adviser Stephen Miller also weighed in after Kirk’s killing, saying that left-wing political organizations constitute “a vast domestic terror movement.”

“We are going to use every resource we have… throughout this government to identify, disrupt, dismantle, and destroy these networks and make America safe again,” Miller said.

But policymakers and the public need reliable evidence and actual data to understand the reality of politically motivated violence. From our research on extremism, it’s clear that the president’s and Miller’s assertions about political violence from the left are not based on actual facts.

Based on our own research and a review of related work, we can confidently say that most domestic terrorists in the US are politically on the right, and right-wing attacks account for the vast majority of fatalities from domestic terrorism.

Political violence rising

The understanding of political violence is complicated by differences in definitions and the recent Department of Justice removal of an important government-sponsored study of domestic terrorists.

Political violence in the US has risen in recent months and takes forms that go unrecognized. During the 2024 election cycle, nearly half of all states reported threats against election workers, including social media death threats, intimidation, and doxing.

Kirk’s assassination illustrates the growing threat. The man charged with the murder, Tyler Robinson, allegedly planned the attack in writing and online.

This follows other politically motivated killings, including the June assassination of Democratic Minnesota state Rep. and former House Speaker Melissa Hortman and her husband.

These incidents reflect a normalization of political violence. Threats and violence are increasingly treated as acceptable for achieving political goals, posing serious risks to democracy and society.

Defining “political violence”

This article relies on some of our research on extremism, other academic research, federal reports, academic datasets, and other monitoring to assess what is known about political violence.

Support for political violence in the US is spreading from extremist fringes into the mainstream, making violent actions seem normal. Threats can move from online rhetoric to actual violence, posing serious risks to democratic practices.

But different agencies and researchers use different definitions of political violence, making comparisons difficult.

Domestic violent extremism is defined by the FBI and Department of Homeland Security as violence or credible threats of violence intended to influence government policy or intimidate civilians for political or ideological purposes. This general framing, which includes diverse activities under a single category, guides investigations and prosecutions. The FBI and DHS do not investigate people in the US for constitutionally protected speech, activism, or ideological beliefs.

Datasets compiled by academic researchers use narrower and more operational definitions. The Global Terrorism Database counts incidents that involve intentional violence with political, social, or religious motivation.

These differences mean that the same incident may or may not appear in a dataset, depending on the rules applied.

The FBI and Department of Homeland Security emphasize that these distinctions are not merely academic. Labeling an event “terrorism” rather than a “hate crime” can change who is responsible for investigating an incident and how many resources they have to investigate it.

For example, a politically motivated shooting might be coded as terrorism in federal reporting, cataloged as political violence by the Armed Conflict Location and Event Data Project, and prosecuted as a homicide or a hate crime at the state level.

Patterns in incidents and fatalities

Despite differences in definitions, several consistent patterns emerge from available evidence.

Politically motivated violence is a small fraction of total violent crime, but its impact is magnified by symbolic targets, timing, and media coverage.

In the first half of 2025, 35 percent of violent events tracked by University of Maryland researchers targeted US government personnel or facilities—more than twice the rate in 2024.

Right-wing extremist violence has been deadlier than left-wing violence in recent years.

Based on government and independent analyses, right-wing extremist violence has been responsible for the overwhelming majority of fatalities, amounting to approximately 75 to 80 percent of US domestic terrorism deaths since 2001.

Illustrative cases include the 2015 Charleston church shooting, when white supremacist Dylann Roof killed nine Black parishioners; the 2018 Tree of Life Synagogue attack in Pittsburgh, where 11 worshippers were murdered; the 2019 El Paso Walmart massacre, in which an anti-immigrant gunman killed 23 people. The 1995 Oklahoma City bombing, an earlier but still notable example, killed 168 in the deadliest domestic terrorist attack in US history.

By contrast, left-wing extremist incidents, including those tied to anarchist or environmental movements, have made up about 10 to 15 percent of incidents and less than 5 percent of fatalities.

Examples include the Animal Liberation Front and Earth Liberation Front arson and vandalism campaigns in the 1990s and 2000s, which were more likely to target property rather than people.

Violence occurred during Seattle May Day protests in 2016, with anarchist groups and other demonstrators clashing with police. The clashes resulted in multiple injuries and arrests. In 2016, five Dallas police officers were murdered by a heavily armed sniper who was targeting white police officers.

Hard to count

There’s another reason it’s hard to account for and characterize certain kinds of political violence and those who perpetrate it.

The US focuses on prosecuting criminal acts rather than formally designating organizations as terrorist, relying on existing statutes such as conspiracy, weapons violations, RICO provisions, and hate crime laws to pursue individuals for specific acts of violence.

Unlike foreign terrorism, the federal government does not have a mechanism to formally charge an individual with domestic terrorism. That makes it difficult to characterize someone as a domestic terrorist.

The State Department’s Foreign Terrorist Organization list applies only to groups outside of the United States. By contrast, US law bars the government from labeling domestic political organizations as terrorist entities because of First Amendment free speech protections.

Rhetoric is not evidence

Without harmonized reporting and uniform definitions, the data will not provide an accurate overview of political violence in the US.

But we can make some important conclusions.

Politically motivated violence in the US is rare compared with overall violent crime. Political violence has a disproportionate impact because even rare incidents can amplify fear, influence policy, and deepen societal polarization.

Right-wing extremist violence has been more frequent and more lethal than left-wing violence. The number of extremist groups is substantial and skewed toward the right, although a count of organizations does not necessarily reflect incidents of violence.

High-profile political violence often brings heightened rhetoric and pressure for sweeping responses. Yet the empirical record shows that political violence remains concentrated within specific movements and networks rather than spread evenly across the ideological spectrum. Distinguishing between rhetoric and evidence is essential for democracy.

Trump and members of his administration are threatening to target whole organizations and movements and the people who work in them with aggressive legal measures—to jail them or scrutinize their favorable tax status. But research shows that the majority of political violence comes from people following right-wing ideologies.

Art Jipson is associate professor of sociology at the University of Dayton, and Paul J. Becker is associate professor of sociology at University of Dayton.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

The Conversation is an independent source of news and views, sourced from the academic and research community. Our team of editors work with these experts to share their knowledge with the wider public. Our aim is to allow for better understanding of current affairs and complex issues, and hopefully improve the quality of public discourse on them.

Right-wing political violence is more frequent, deadly than left-wing violence Read More »

China blocks sale of Nvidia AI chips

AI, china, NVIDIA, Policy, syndication, trump tariffs / Mike M. / September 17, 2025

“The message is now loud and clear,” said an executive at one of the tech companies. “Earlier, people had hopes of renewed Nvidia supply if the geopolitical situation improves. Now it’s all hands on deck to build the domestic system.”

Nvidia started producing chips tailored for the Chinese market after former US President Joe Biden banned the company from exporting its most powerful products to China, in an effort to rein in Beijing’s progress on AI.

Beijing’s regulators have recently summoned domestic chipmakers such as Huawei and Cambricon, as well as Alibaba and search engine giant Baidu, which also make their own semiconductors, to report how their products compare against Nvidia’s China chips, according to one of the people with knowledge of the matter.

They concluded that China’s AI processors had reached a level comparable to or exceeding that of the Nvidia products allowed under export controls, the person added.

The Financial Times reported last month that China’s chipmakers were seeking to triple the country’s total output of AI processors next year.

“The top-level consensus now is there’s going to be enough domestic supply to meet demand without having to buy Nvidia chips,” said an industry insider.

Nvidia introduced the RTX Pro 6000D in July during Huang’s visit to Beijing, when the US company also said Washington was easing its previous ban on the H20 chip.

China’s regulators, including the CAC, have warned tech companies against buying Nvidia’s H20, asking them to justify having purchased them over domestic products, the FT reported last month.

The RTX Pro 6000D, which the company has said could be used in automated manufacturing, was the last product Nvidia was allowed to sell in China in significant volumes.

Alibaba, ByteDance, the CAC, and Nvidia did not immediately respond to requests for comment.

Additional reporting by Eleanor Olcott in Zhengzhou.

China blocks sale of Nvidia AI chips Read More »

“China keeps the algorithm”: Critics attack Trump’s TikTok deal

bytedance, Donald Trump, national security, Policy, tiktok / Mike M. / September 17, 2025

However, Trump seems to think that longtime TikTok partner Oracle taking a bigger stake while handling Americans’ user data at its facilities in Texas will be enough to prevent remaining China-based owners—which will maintain less than a 20 percent stake—from allegedly spying, launching disinformation campaigns, or spreading other kinds of propaganda.

China previously was resistant to a forced sale of TikTok, FT reported, even going so far as to place export controls on algorithms to keep the most lucrative part of TikTok in the country. And “it remains unclear to what extent TikTok’s Chinese parent would retain control of the algorithm in the US as part of a licensing deal,” FT noted.

On Tuesday, Wang Jingtao, deputy head of China’s cyber security regulator, did not go into any detail on how China’s access to US user data would be restricted under the deal. Instead, Wang only noted that ByteDance would “entrust the operation of TikTok’s US user data and content security,” presumably to US owners, FT reported.

One Asia-based investor told FT that the US would use “at least part of the Chinese algorithm” but train it on US user data, while a US advisor accused Trump of chickening out and accepting a deal that didn’t force a sale of the algorithm.

“After all this, China keeps the algorithm,” the US advisor said.

To the Asia-based investor, it seemed like Trump gave China exactly what it wants, since “Beijing wants to be seen as exporting Chinese technology to the US and the world.”

It’s likely more details will be announced once Trump and Chinese President Xi Jinping hold a phone conference on Friday. ByteDance has yet to comment on the deal and did not respond to Ars’ request to comment.

“China keeps the algorithm”: Critics attack Trump’s TikTok deal Read More »

A new report finds China’s space program will soon equal that of the US

artemis, china, redshift, Space, space program / Mike M. / September 16, 2025

As Jonathan Roll neared completion of a master’s degree in science and technology policy at Arizona State University three years ago, he did some research into recent developments by China’s ascendant space program. He came away impressed by the country’s growing ambitions.

Now a full-time research analyst at the university, Roll was recently asked to take a deeper dive into Chinese space plans.

“I thought I had a pretty good read on this when I was finishing grad school,” Roll told Ars. “That almost everything needed to be updated, or had changed three years later, was pretty scary. On all these fronts, they’ve made pretty significant progress. They are taking all of the cues from our Western system about what’s really galvanized innovation, and they are off to the races with it.”

Roll is the co-author of a new report, titled “Redshift,” on the acceleration of China’s commercial and civil space activities, and the threat these pose to similar efforts in the United States. Published on Tuesday, the report was sponsored by the US-based Commercial Space Federation, which advocates for the country’s commercial space industry. It is a sobering read, and comes as China not only projects to land humans on the lunar surface before the US can return, but is advancing across several spaceflight fronts to challenge America.

“The trend line is unmistakable,” the report states. “China is not only racing to catch up—it is setting pace, deregulating, and, at times, redefining what leadership looks like on and above Earth. This new space race will not be won with a single breakthrough or headline achievement, but with sustained commitment, clear-eyed vigilance, and a willingness to adapt over decades.”

A new report finds China’s space program will soon equal that of the US Read More »

How Nissan leveraged its driver assist to cut traffic jams

Cars, Nissan, partially automated driving, traffic congestion / Mike M. / September 16, 2025

Instead, CCM works by having a lead car, or “probe,” send information to following CCM-equipped cars, which are separated by non-CCM cars between them. The information from the probe car lets the following cars keep an appropriate distance from each other—between 30 and 60 seconds—and if there’s a slowdown ahead, the following cars will decelerate more gently over time, preventing the kind of concertina action that triggers traffic jams when human drivers see someone slowing down in front of them.

Jerry Chou, a senior researcher at Nissan’s Silicon Valley center, described CCM to me as “mixed autonomy—that means a mix of the controlled vehicles and other human driven vehicles in between.” Instead of DSRC, the cars use their embedded LTE modems to communicate via Nissan’s cloud.

As most people who have used adaptive cruise control know, if your following distance is too large, other drivers will often cut in, causing you to decelerate. “So we did spend some time to balance this phenomenon and the performance of our system. So there’s some parameters we continue to control to balance this,” Chou told me.

The view from inside a Nissan Ariya equipped with an experimental congestion management system. — Note the test equipment that’s fitted to the dash of this CCM-equipped Nissan Ariya test vehicle. Credit: Nissan

Next, I asked Chou what percentage of cars in traffic would need to be CCM-enabled to effect a reduction in congestion?

“So in our simulations we tried different penetration rates… and we saw that our benefits increase proportionally to penetration rates. But we already can see some good results at around 4–5 percent penetration,” Chou told me. “But you know, that’s actually one challenge of experimental. Since our experiment only has a few cars, we have been thinking about how to control just these few cars to see some results.”

Future refinements for the system include giving the humans some feedback on why their cars are slowing (in part so they don’t countermand the system and just accelerate manually). If that proves successful, we may even see CCM licensed to other automakers in the future, Chou said.

How Nissan leveraged its driver assist to cut traffic jams Read More »

Millions turn to AI chatbots for spiritual guidance and confession

AI, AI assistants, AI behavior, AI sycophancy, Biz & IT, chatbots, chatgpt, Gemini, generative ai, Google, large language models, machine learning, openai / Mike M. / September 16, 2025

Privacy concerns compound these issues. “I wonder if there isn’t a larger danger in pouring your heart out to a chatbot,” Catholic priest Fr. Mike Schmitz told The Times. “Is it at some point going to become accessible to other people?” Users share intimate spiritual moments that now exist as data points in corporate servers.

Some users prefer the chatbots’ non-judgmental responses to human religious communities. Delphine Collins, a 43-year-old Detroit preschool teacher, told the Times she found more support on Bible Chat than at her church after sharing her health struggles. “People stopped talking to me. It was horrible.”

App creators maintain that their products supplement rather than replace human spiritual connection, and the apps arrive as approximately 40 million people have left US churches in recent decades. “They aren’t going to church like they used to,” Beck said. “But it’s not that they’re less inclined to find spiritual nourishment. It’s just that they do it through different modes.”

Different modes indeed. What faith-seeking users may not realize is that each chatbot response emerges fresh from the prompt you provide, with no permanent thread connecting one instance to the next beyond a rolling history of the present conversation and what might be stored as a “memory” in a separate system. When a religious chatbot says, “I’ll pray for you,” the simulated “I” making that promise ceases to exist the moment the response completes. There’s no persistent identity to provide ongoing spiritual guidance, and no memory of your spiritual journey beyond what gets fed back into the prompt with every query.

But this is spirituality we’re talking about, and despite technical realities, many people will believe that the chatbots can give them divine guidance. In matters of faith, contradictory evidence rarely shakes a strong belief once it takes hold, whether that faith is placed in the divine or in what are essentially voices emanating from a roll of loaded dice. For many, there may not be much difference.

Millions turn to AI chatbots for spiritual guidance and confession Read More »

Modder injects AI dialogue into 2002’s Animal Crossing using memory hack

AI, AI integration, Animal Crossing, Biz & IT, chatgpt, Dolphin emulator, Emulation, game hacking, game modding, GameCube, gaming, generative ai, Google, large language models, machine learning, Nintendo, openai, retrogaming, retrotech, Tom Nook / Mike M. / September 13, 2025

But discovering the addresses was only half the problem. When you talk to a villager in Animal Crossing, the game normally displays dialogue instantly. Calling an AI model over the Internet takes several seconds. Willison examined the code and found Fonseca’s solution: a watch_dialogue() function that polls memory 10 times per second. When it detects a conversation starting, it immediately writes placeholder text: three dots with hidden pause commands between them, followed by a “Press A to continue” prompt.

“So the user gets a ‘press A to continue’ button and hopefully the LLM has finished by the time they press that button,” Willison noted in a Hacker News comment. While players watch dots appear and reach for the A button, the mod races to get a response from the AI model and translate it into the game’s dialog format.

Learning the game’s secret language

Simply writing text to memory froze the game. Animal Crossing uses an encoded format with control codes that manage everything from text color to character emotions. A special prefix byte (0x7F) signals commands rather than characters. Without the proper end-of-conversation control code, the game waits forever.

“Think of it like HTML,” Fonseca explains. “Your browser doesn’t just display words; it interprets tags … to make text bold.” The decompilation community had documented these codes, allowing Fonseca to build encoder and decoder tools that translate between a human-readable format and the GameCube’s expected byte sequences.

A screenshot of LLM-powered dialog injected into Animal Crossing for the GameCube. — A screenshot of LLM-powered dialog injected into *Animal Crossing* for the GameCube. Credit: Joshua Fonseca

Initially, he tried using a single AI model to handle both creative writing and technical formatting. “The results were a mess,” he notes. “The AI was trying to be a creative writer and a technical programmer simultaneously and was bad at both.”

The solution: split the work between two models. A Writer AI creates dialogue using character sheets scraped from the Animal Crossing fan wiki. A Director AI then adds technical elements, including pauses, color changes, character expressions, and sound effects.

The code is available on GitHub, though Fonseca warns it contains known bugs and has only been tested on macOS. The mod requires Python 3.8+, API keys for either Google Gemini or OpenAI, and Dolphin emulator. Have fun sticking it to the man—or the raccoon, as the case may be.

Modder injects AI dialogue into 2002’s Animal Crossing using memory hack Read More »

California bill lets renters escape exclusive deals between ISPs and landlords

cable lobby, FCC, fcc bulk billing, Policy / Mike M. / September 13, 2025

Opt-out from bulk billing

Bill author says law “gives this industry an opportunity to treat people fairly.”

Credit: Getty Images | Yuichiro Chino

California’s legislature this week approved a bill to let renters opt out of bulk-billing arrangements that force them to pay for Internet service from a specific provider.

The bill says that by January 1, a landlord must “allow the tenant to opt out of paying for any subscription from a third-party Internet service provider, such as through a bulk-billing arrangement, to provide service for wired Internet, cellular, or satellite service that is offered in connection with the tenancy.” If a landlord fails to do so, the tenant “may deduct the cost of the subscription to the third-party Internet service provider from the rent,” and the landlord would be prohibited from retaliating.

The bill passed the state Senate in a 30–7 vote on Wednesday but needs Gov. Gavin Newsom’s signature to become law. It was approved by the state Assembly in a 75–0 vote in April.

Assemblymember Rhodesia Ransom, a Democratic lawmaker who authored the bill, told Ars today that lobby groups for Internet providers and real estate companies have been “working really hard” to defeat it. But she expects Newsom will approve.

“I strongly believe that the governor is going to look at what this bill provides as far as protections for tenants and sign it into law,” Ransom said in a phone interview.

“Just treat people fairly”

Ransom disputed claims from lobby groups that bulk billing reduces Internet prices for tenants.

“This is kind of like a first step in trying to give this industry an opportunity to just treat people fairly. It’s not super restrictive. We are not banning bulk billing. We’re not even limiting how much money the people can make. What we’re saying here with this bill is that if a tenant wants to opt out of the arrangement, they should be allowed to opt out,” she said.

A stricter bill could have told landlords that “you can’t charge the customer more than you’re paying. We could have put a cap on the amount that you’re able to charge,” she said. “There’s so many other things that we could have done that would’ve been a lot less business-friendly. But the goal was not to harm business, the goal was to help people.”

In theory, bulk billing could reduce prices for tenants if discounts negotiated between landlords and Internet providers were passed on to renters. But, Ransom said, “where there would be an opportunity for these huge discounts to be passed on to tenants, it’s not happening. We know of thousands of tenants across the state who are in landlord-tenant agreements where the landlord is actually adding an additional bonus for themselves, pocketing change, and not passing the discount on to the tenants… once we started working on this bill, we started to hear more and more about places where people were stuck in these agreements and their landlords were not letting them out.”

Ransom said not all landlords do this and that it is generally “the large corporate landlords” who own hundreds or thousands of properties that “were the ones who were reluctant to let their tenants out.”

State bill similar to abandoned FCC plan

California’s action comes about eight months after the Federal Communications Commission abandoned a proposal to give tenants the right to opt out of bulk billing for Internet service. The potential federal action was proposed in March 2024 by then-FCC Chairwoman Jessica Rosenworcel, but nixed in January 2025 by Chairman Brendan Carr.

Bulk billing contracts are only banned by the FCC when they give a provider the exclusive right to access and serve a building. Despite that restriction, a bulk billing deal between an ISP and landlord can make it less financially feasible for other providers to serve a multi-unit building. Letting people opt out of bulk billing arrangements makes serving a building at least slightly more viable for a competing provider.

Ransom said the FCC action “was very unfortunate” and “give[s] a disadvantage to people who are already at the mercy of landlords.”

Cable lobby calls it an “anti-affordability bill”

The California bill was not welcomed by lobby groups for Internet providers and landlords. The California Broadband & Video Association, which represents cable companies, paid for a sponsored commentary in several news publications to express its opposition.

“AB 1414 is an anti-affordability bill masked as consumer protection, and it will only serve to widen the digital divide in California,” wrote the lobby group’s CEO, Janus Norman.

Norman complained that property owners would have “to provide a refund to tenants who decline the Internet service provided through the building’s contract with a specific Internet service provider.” He argued that without bulk billing, “low-income families and tenants risk losing access altogether.”

Letting tenants opt out of bulk deals “undermines the basis of the cost savings and will lead to bulk billing being phased out,” Norman wrote. This “will result in higher bills for everyone, including those already struggling,” he claimed.

“The truth, very simply, is this: bulk billing is good for consumers,” the cable industry commentary said. “Taking away bulk discounts raises total housing costs when Californians can least afford it.”

The bill also drew opposition from the Real Estate Technology & Transformation Center (RETTC). The group’s sponsors include real estate companies and Internet providers AT&T, Comcast, and Cox. Another notable sponsor of RETTC is RealPage, which has faced claims from the US government and state attorneys general that its software distorts competition in rental housing by helping landlords collectively set prices.

“AB 1414 introduces an opt-out requirement that would fundamentally undermine the economics of bulk billing,” the RETTC said. “By fragmenting service, it could destabilize networks and reduce the benefits residents and operators rely on today.” The group claimed the bill could lead to “higher broadband costs for renters, reduced ISP investment in multifamily housing, disruption of property-wide smart technology, [and] widening of the digital divide in California.”

The RETTC said it joined with the National Apartment Association and the California Rental Housing Association to detail the groups’ concerns directly to the bill sponsors.

Wireless providers could get a boost

The California Broadband & Video Association seems to be worried about wireless providers serving buildings wired up with cable. The group’s commentary claimed that “the bill’s lack of technology neutrality also creates winners and losers, granting certain types of providers an unfair advantage over their competitors.”

Ransom said her bill may be especially helpful for wireless or satellite providers because they wouldn’t need to install wires in each building.

“This does help with market competition, and in fact some of our support came from some of the smaller Internet service providers… and because this bill is technology-neutral, it helps with not only the current technology, but any new technology that comes out,” she said.

While Ransom’s bill could help make broadband more affordable for renters, California lawmakers recently abandoned a more aggressive effort to require affordable broadband plans. Assemblymember Tasha Boerner proposed a state law that would force Internet service providers to offer $15 monthly plans to people with low incomes but tabled the bill after the Trump administration threatened to block funding for expanding broadband networks.

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.

California bill lets renters escape exclusive deals between ISPs and landlords Read More »

Microsoft dodges EU fine by unbundling Teams from Office

competition, EU, microsoft, Microsoft Teams, Policy, syndication / Mike M. / September 12, 2025

Microsoft has avoided an EU fine after the US tech group offered concessions on how it packages together its Teams and Office products, ending a long-running antitrust investigation by the bloc’s regulators.

The probe, which began after a 2020 complaint from Slack, now part of Salesforce, accused Microsoft of abusing its market dominance by tying its video conferencing tool to its widely used suite of productivity applications.

Since the initial complaint, Microsoft has unbundled Teams from Office 365 in the EU, but critics said the changes were too narrow.

In May, the $3.7 trillion software giant promised concessions, such as continuing the Teams and Office separation for seven years.

After a market test, Microsoft has since made additional commitments, such as publishing more information on so-called “interoperability” or the ability to use its products with others made by rivals.

These new pledges have satisfied the EU’s regulator, which said on Friday that it helped to restore fair competition and open the market to other providers.

Microsoft dodges EU fine by unbundling Teams from Office Read More »

Pentagon begins deploying new satellite network to link sensors with shooters

falcon 9, launch, military space, Science, Space, space development agency, spacex, US Space Force, Vandenberg space force base, york space systems / Mike M. / September 11, 2025

“This is the first time we’ll have a space layer fully integrated into our warfighting operations.”

A SpaceX Falcon 9 rocket lifts off from Vandenberg Space Force Base, California, with a payload of 21 data-relay satellites for the US military’s Space Development Agency. Credit: SpaceX

The first 21 satellites in a constellation that could become a cornerstone for the Pentagon’s Golden Dome missile-defense shield successfully launched from California Wednesday aboard a SpaceX Falcon 9 rocket.

The Falcon 9 took off from Vandenberg Space Force Base, California, at 7: 12 am PDT (10: 12 am EDT; 14: 12 UTC) and headed south over the Pacific Ocean, heading for an orbit over the poles before releasing the 21 military-owned satellites to begin several weeks of activations and checkouts.

These 21 satellites will boost themselves to a final orbit at an altitude of roughly 600 miles (1,000 kilometers). The Pentagon plans to launch 133 more satellites over the next nine months to complete the build-out of the Space Development Agency’s first-generation, or Tranche 1, constellation of missile-tracking and data-relay satellites.

“We had a great launch today for the Space Development Agency, putting this array of space vehicles into orbit in support of their revolutionary new architecture,” said Col. Ryan Hiserote, system program director for the Space Force’s assured access to space launch execution division.

Over the horizon

Military officials have worked for six years to reach this moment. The Space Development Agency (SDA) was established during the first Trump administration, which made plans for an initial set of demonstration satellites that launched a couple of years ago. In 2022, the Pentagon awarded contracts for the first 154 operational spacecraft. The first batch of 21 data-relay satellites built by Colorado-based York Space Systems is what went up Wednesday.

“Back in 2019, when the SDA was stood up, it was to do two things. One was to make sure that we can do beyond line of sight targeting, and the other was to pace the threat, the emerging threat, in the missile-warning and missile-tracking domain. That’s what the focus has been,” said Gurpartap “GP” Sandhoo, the SDA’s acting director.

Secretary of the Air Force Troy Meink and Sen. Kevin Cramer (R-N.D.) pose with industry and government teams in front of the Space Development’s first 21 operational satellites at Vandenberg Space Force Base, California. Cramer is one the most prominent backers of the Golden Dome program in the US Senate. Credit: US Air Force/Staff Sgt. Daekwon Stith

Historically, the military communications and missile-warning networks have used a handful of large, expensive satellites in geosynchronous orbit some 22,000 miles (36,000 kilometers) above the Earth. This architecture was devised during the Cold War and is optimized for nuclear conflict and intercontinental ballistic missiles.

For example, the military’s ultra-hardened Advanced Extremely High Frequency satellites in geosynchronous orbit are designed to operate through an electromagnetic pulse and nuclear scintillation. The Space Force’s missile-warning satellites are also in geosynchronous orbit, with infrared sensors tuned to detect the heat plume of a missile launch.

The problem? Those satellites cost more than $1 billion a pop. They’re also vulnerable to attack from a foreign adversary. Pentagon officials say the SDA’s satellite constellation, officially called the Proliferated Warfighter Space Architecture, is tailored to detect and track more modern threats, such as smaller missiles and hypersonic weapons carrying conventional warheads. It’s easier for these missiles to evade the eyes of older early warning satellites.

What’s more, the SDA’s fleet in low-Earth orbit will have numerous satellites. Losing one or several satellites to an attack would not degrade the constellation’s overall capability. The SDA’s new relay satellites cost between $14 and $15 million each, according to Sandhoo. The total cost of the first tranche of 154 operational satellites totals approximately $3.1 billion.

Multi-mission satellites

These satellites will not only detect and track ballistic and hypersonic missile launches; they will also transmit signals between US forces using an existing encrypted tactical data link network known as Link 16. This UHF system is used by NATO and other US allies to allow military aircraft, ships, and land forces to share tactical information through text messages, pictures, data, and voice communication in near real time, according to the SDA’s website.

Up to now, Link 16 radios were ubiquitous on fighter jets, helicopters, naval vessels, and missile batteries. But they had a severe limitation. Link 16 was only able to close a radio link with a clear line of sight. The Space Development Agency’s satellites will change that, providing direct-to-weapon connectivity from sensors to shooters on Earth’s surface, in the air, and in space.

The relay satellites, which the SDA calls the transport layer, are also equipped with Ka-band and laser communication terminals for higher-bandwidth connectivity.

“What the transport layer does is it extends beyond the line of sight,” Sandhoo said. “Now, you’re able to talk not only to within a couple of miles with your Link 16 radios, (but) we can use space to, let’s say, go from Hawaii out to Guam using those tactical radios, using a space layer.”

The Space Development Agency’s “Tranche 1” architecture includes 154 operational satellites, 126 for data relay and 28 for missile tracking. With this illustration, the SDA does its best to show how the complex architecture is supposed to work. Credit: Space Development Agency

Another batch of SDA relay satellites will launch next month, and more will head to space in November. In all, it will take 10 launches to fully deploy the SDA’s Tranche 1 constellation. Six of those missions will carry data-relay satellites, and four will carry satellites with sensors to detect and track missile launches. The Pentagon selected several contractors to build the satellites, so the military is not reliant on a single company. The builders of the SDA’s operational satellites include York, Lockheed Martin, Northrop Grumman, and L3Harris.

“We will increase coverage as we get the rest of those launches on orbit,” said Michael Eppolito, the SDA’s acting deputy director.

The satellites will connect with one another using inter-satellite laser links, creating a mesh network with sufficient range to provide regional communications, missile warning, and targeting coverage over the Western Pacific beginning in 2027. US Indo-Pacific Command, which oversees military operations in this region, is slated to become the first combatant command to take up use of the SDA’s satellite constellation.

This is not incidental. US officials see China as the nation’s primary strategic threat, and Indo-Pacific Command would be on the front lines of any future conflict between Chinese and US forces. The SDA has contracts in place for more than 270 second-generation, or Tranche 2, satellites, to further expand the network’s reach. There’s also a third generation in the works, but the Pentagon has paused part of the SDA’s Tranche 3 program to evaluate other architectures, including one offered by SpaceX.

Teaching tactical operators to use the new capabilities offered by the SDA’s satellite fleet could be just as challenging as building the network itself. To do this, the Pentagon plans to put soldiers, sailors, airmen, and marines through “warfighter immersion” training beginning next year. This training will allow US forces to “get used to using space from this construct,” Sandhoo said.

“This is different than how it has been done in the past,” Sandhoo said. “This is the first time we’ll have a space layer actually fully integrated into our warfighting operations.”

The SDA’s satellite architecture is a harbinger for what’s to come with the Pentagon’s Golden Dome system, a missile-defense shield for the US homeland proposed by President Donald Trump in an executive order in January. Congress authorized a down payment on Golden Dome in July, the first piece of funding for what the White House says will cost $175 billion over the next three years.

Golden Dome, as currently envisioned, will require thousands of satellites in low-Earth orbit to track missile launches and space-based interceptors to attempt to shoot them down. The Trump administration hasn’t said how much of the shield might be deployed by the end of 2028, or what the entire system might eventually cost.

But the capabilities of the SDA’s satellites will lay the foundation for any regional or national missile-defense shield. Therefore, it seems likely that the military will incorporate the SDA network into Golden Dome, which, at least at first, is likely to consist of technologies already in space or nearing launch. Apart from the Space Development Agency’s architecture in low-Earth orbit (LEO), the Space Force was already developing a new generation of missile-warning satellites to replace aging platforms in geosynchronous orbit (GEO), plus a fleet of missile-warning satellites to fly at a midrange altitude between LEO and GEO.

Air Force Gen. Gregory Guillot, commander of US Northern Command, said in April that Golden Dome “for the first time integrates multiple layers into one system that allows us to detect, track, and defeat multiple types of threats that affect us in different domains.

“So, while a lot of the components and the requirements were there in the past, this is the first time that it’s all tied together in one system,” he said.

Stephen Clark is a space reporter at Ars Technica, covering private space companies and the world’s space agencies. Stephen writes about the nexus of technology, science, policy, and business on and off the planet.

Pentagon begins deploying new satellite network to link sensors with shooters Read More »

AI #133: America Could Use More Energy

America / Mike M. / September 11, 2025

Even in quiet weeks like this one, there are noticeable incremental upgrades. The cost of the best video generation tool, Veo 3, went down by half. ChatGPT now offers conversation branching. Claude can directly edit files. Yet it is a good time to ask about the missing results. Where are all the AI agents? If AI coding is so good why aren’t we seeing a surge in GitHub repositories or iPhone apps?

A lot of the focus since the rollout of GPT-5 remains on the perception and policy fronts, especially regarding views of AI progress. The botched rollout of what is ultimately a very good (but not mind blowing) model gave a lot of people the wrong impression, so I have to remind everyone that once again that AI continues to make rapid progress. Meanwhile, we must also notice that OpenAI’s actions in the public sphere have once again because appreciably worse, as they descend into paranoia and bad faith lobbying, including baseless legal attacks on nonprofits.

Also this week: Yes, AI Continues To Make Rapid Progress, Including Towards AGI and OpenAI #14: OpenAI Descends Into Paranoia And Bad Faith Lobbying.

Language Models Offer Mundane Utility. Use AI to simulate a simulacra?
Productivity Puzzles. Where is the massive flood of additional software?
Language Models Don’t Offer Mundane Utility. Why no progress on GPTs?
Huh, Upgrades. Claude can edit files, ChatGPT can branch, Veo 3 50% off.
On Your Marks. ClockBench? AIs remain remarkably bad at this one.
Choose Your Fighter. Karpathy likes GPT-5-Pro, GPT-5 lacks metacognition.
Fun With Media Generation. AI assisted $30 million animated feature is coming.
Deepfaketown and Botpocalypse Soon. Dead Internet Theory, what a surprise.
Unprompted Attention. No, a prompt cannot entirely halt hallucinations.
Get My Agent On The Line. Where are all the useful AI agents?
They Took Our Jobs. I never thought the leopards would automate MY job.
A Young Lady’s Illustrated Primer. We built this system on proof of work.
Levels of Friction. When detection costs drop dramatically, equilibria break.
The Art of the Jailbreak. AI, let me talk to your manager.
Get Involved. Anthropic safety fellows program head, Foresight Institute.
Introducing. We could all use a friend, but not like this.
In Other AI News. EBay, novel math, Anthropic enforces bans and more.
Show Me the Money. Valuations up enough Anthropic can pay out $1.5 billion.
Quiet Speculations. Speeding up your releases versus speeding up your progress.
The Quest for Sane Regulations. Anthropic endorses SB 53, as do I.
Chip City. Nvidia loves selling to China, Department of Energy hates energy.
The Week in Audio. Me, Truell, Nanda, Altman on Carlson, Bell and Ruiz.
All Words We Choose Shall Lose All Meaning. It is the curse we must accept.
Hunger Strike. If you believed that, why wouldn’t you? Oh, you did.
Rhetorical Innovation. Nvidia continues calling everyone they dislike ‘doomer.’
Misaligned! Might want to keep an eye on those suggested changes.
Hallucinations. We can greatly reduce hallucinations if we care enough.
Aligning a Smarter Than Human Intelligence is Difficult. Janus explains.
The Lighter Side. It’s going to take a while to get this far.

Reese Witherspoon, in what is otherwise a mediocre group puff piece about The Morning Show, talks about her use of AI. She uses Perplexity and Vetted AI (a shopping assistant I hadn’t heard of) and Simple AI which makes phone calls to businesses for you. I am skeptical that Vetted is ever better than using ChatGPT or Claude, and I haven’t otherwise heard of people having success with Simple AI or similar services, but I presume it’s working for her.

She also offers this quote:

Reese Witherspoon: It’s so, so important that women are involved in AI…because it will be the future of filmmaking. And you can be sad and lament it all you want, but the change is here. It will never be a lack of creativity and ingenuity and actual physical manual building of things. It might diminish, but it’s always going to be the highest importance in art and in expression of self.

The future of filmmaking certainly involves heavy use of AI in various ways. That’s baked in. The mistake here is in assuming there won’t be even more change, as Reese like most people isn’t yet feeling the AGI or thinking ahead to what it would mean.

AI can simulate a ‘team of rivals’ that can then engage in debate. I’ve never been drawn to that as a plan but it doesn’t seem crazy.

Can we use AI to simulate human behavior realistically enough to conduct sociological experiments? Benjamin Manning and John Horton give it a shot with a paper where they have the AI play ‘a highly heterogeneous population of 883,320 novel games.’ In preregistered experiments, AI agents constructed using seed data then on related but distinct games predict human behavior better than either out-of-the-box agents or game-theoretic equilibria.

That leaves out other obvious things you could try in order to get a better distribution. They try some things, but they don’t try things like actively asking it to predict the distribution of answers humans would give.

They use as their example the ‘11-20 money game’ where you request some number of dollars from 11 to 20, and you get that many dollars, plus an extra $20 if the other player requested one dollar more than you did.

If you simply ask an LLM to play, you get a highly inhuman distribution, here is GPT-4o doubling down on 19:

That’s not a crazy distribution for a particular human, I’m sure many people mostly choose 19, nor is it too obviously a terrible strategy. Given actual human behavior there is a right answer. Maybe 20 is too common among humans, even though it seems like an obvious mistake against any realistic distribution. My instinct if I got to play this once against humans was to answer 18, but I realized after saying that I was probably being anchored by GPT-4o’s answer. I do think that any answer outside 16-18 is clearly a mistake versus humans unless you think a lot of them will misunderstand the game and thereby choose 20.

GPT-5-Pro predicted performance well when I asked it to, but then I realized it was looking at prior research on this game on the web, so that doesn’t count, and it might well be in the training data.

Benjamin Manning: For the 11-20 money request game, the theory is level-k thinking, and the seed game is the human responses from the original paper. We construct a set of candidate agents based on a model of level-k thinking and then optimize them to match human responses with high accuracy.

When we have these optimized agents play two new related, but distinct games, the optimized set performs well in matching these out-of-sample human distributions. The off-the-shelf LLM still performs poorly.

…

We then put the strategic and optimized agents to an extreme test. We created a population of 800K+ novel strategic games, sampled 1500, which the agents then played in 300,000 simulations. But we first have 3 humans (4500 total) play each game in a pre-registered experiment.

Optimized agents predict the human responses far better than an off-the-shelf baseline LLM (3x) and relevant game-theoretic equilibria (2x). In 86% of the games, all human subjects chose a strategy in support of the LLM simulations; only 18% were in support of the equilibria.

I find this all great fun and of theoretical interest, but in terms of creating useful findings I am far more skeptical. Making simulated predictions in these toy economic games is too many levels removed from what we want to know.

Arnold Kling shares his method for using AI to read nonfiction books, having AI summarize key themes, put them into his own words, get confirmation he is right, get examples and so on. He calls it ‘stop, look and listen.’

Arnold Kling: Often, what you remember about a book can be reduced to a tweet, or just a bumper sticker. So when I’ve finished a half-hour conversation with an AI about a book, if I have a solid handle on five key points, I am ahead of the game.

My question is, isn’t that Kling’s method of not reading a book? Which is fine, if you are reading the type of book where 90% or more of it is fluff or repetition. It does question why you are engaging with a book like that in the first place.

Is the book ‘written for the AIs’? With notably rare exceptions, not yet.

Is the book an expansion of a 1-10 page explanation, or even a sentence or two, that is valuable but requires repeated hammering to get people to listen? Does it require that one ‘bring the receipts’ in some form so we know they exist and can check them? Those are much more likely, but I feel we can do better than doing our own distillations, even the AI version.

Thus, I read few books, and try hard to ensure the ones I do read are dense. If I’m going to bother reading a non-fiction book, half or more of the time I’m eyeing a detailed review.

Here’s another counterpoint.

Samo Burja: I have no idea why people would summarize books through AI. When the right time comes for a book, every sentence gives new generative ideas and connections. Why not have the AI eat for you too?

It’s 2025. No one’s job or even education really requires people to pretend to have this experience through reading entire books. Reading has been liberated as pure intellectual generation. Why then rob yourself of it?

The whole objection from Kling is that most books don’t offer new generative ideas and connections in every sentence. Even the Tyler Cowen rave is something like ‘new ideas on virtually every page’ (at the link about Keynes) which indicates that the book is historically exceptional. I do agree with the conclusion that you don’t want to rob yourself of the reading when the reading is good enough, but the bar is high.

Also while we’re talking about books, or why for so many it’s been Moby Dick summer:

Dwarkesh Patel: I find it frustrating that almost every nonfiction book is basically just a history lesson, even if it’s nominally about some science/tech/policy topic.

Nobody will just explain how something works.

Books about the semiconductor industry will never actually explain the basic process flow inside a fab, but you can bet that there will be a minute-by-minute recounting of a dramatic 1980s Intel boardroom battle.

Dan Hendrycks: Agreed. Even if someone tries to be intellectually incisive and not chatty, they usually can’t outcompete a textbook.

An exception you read it for the author’s lens on the world (e.g., Antifragile).

Iterate having AIs produce and validate encounters for a role playing game.

Gemini is very good at analyzing your golf swing and explaining how to fix it, at least at beginner levels.

Mike Judge fails to notice AI speeding up his software development in randomized tests, as he attempted to replicate the METR experiment that failed to discover speedups in experts working on their own code bases. Indeed, he found a 21% slowdown, similar to the METR result, although it is not statistically significant.

Arnold Kling reasonably presumes this means Judge is probably at least 98th percentile for developers, and that his experience was the speedup was dramatic. Judge definitely asserts far too much when he says the tools like Cursor ‘don’t work for anyone.’ I can personally say that I am 100% confident they work for me, as in the tasks I did using Cursor would have been impossible for me to do on my own in any reasonable time frame.

But Judge actually has a strong argument we don’t reckon with enough. If AI is so great, where is the shovelware, where are the endless Tetris clones and what not? Instead the number of new apps isn’t changing on iOS, and if anything is falling on Android, there’s no growth in Steam releases or GitHub repositories.

This is indeed highly weird data. I know that AI coding increased my number of GitHub repos from 0 to 1, but that isn’t looking widespread. Why is this dog not barking in the nighttime?

I don’t know. It’s a good question. It is very, very obvious that AI when used well greatly improves coding speed and ability, and there’s rapidly growing use. Thoughts?

One place we do see this kind of explosion is patents.

Rohan Paul: US Patent exploding with AI revolution.

This looks like a small number of patents, then with the internet a huge jump, then with AI another big jump where the graph is going vertical but hasn’t gone up that much yet. GPT-5-Pro estimates about half the increase is patents is inventions related to AI, and about half is due to easier filing, including for clearing backlogs.

That would mean this doesn’t yet represent a change in the rate of real inventions outside of AI. That illustrates why I have long been skeptical of ‘number of patents’ as a measure, for pretty much all purposes, especially things like national comparisons or ranking universities or companies. It is also a lagging indicator.

Why no progress here either?

Ethan Mollick: I’ll note again that it seems nuts that, despite every AI lab launching a half-dozen new products, nobody is doing anything with GPTs, including OpenAI.

When I talk to people at companies, this is still the way non-technical people share prompts on teams. No big change in 2 years.

Its fine if it turns out that GPTs/Gems/whatever aren’t the future, but it seems reasonably urgent to roll out something else that makes sharing prompts useful across teams and organizations. Prompt libraries are still important, and they are still awkward cut-and-paste things.

GPTs seem like inferior versions of projects in many ways? The primary virtual-GPT I use is technically a project. But yes, problems of this type seem like high value places to make progress and almost no progress is being made.

WSJ reports that many consultants promising to help with AI overpromise and underdeliver while essentially trying to learn AI on the job and the client’s dime, as spending on AI-related consulting triples to $3.75 billion in 2024, I am shocked, shocked to find that going on in this establishment. Given the oversized payoffs when such consulting works, if they didn’t often underdeliver then they’re not being used enough.

Meanwhile McKinsey is scrambling to pivot to AI agents as it realizes that AI will quickly be able to do most of what McKinsey does. For now it’s fine, as AI and related technology now makes up 40% of their revenue.

Claude can now directly create and edit files such as Excel spreadsheets, documents, PowerPoint slide decks and PDFs, if you enable it under experimental settings. They can be saved directly to Google Drive.

ChatGPT adds full support for MCP tools.

Veo 3 and Veo 3 fast cut prices and join the Gemini API, and they are adding support for 9: 16 vertical and 1080 HD outputs.

Google AI Developers: The new pricing structure is effective immediately:

🔹 Veo 3 $0.40 /sec (from $0.75)

🔹 Veo 3 Fast $0.15/sec (from $0.40)

ChatGPT now has an option to branch a conversation, from any point, into a new chat.

This is a big deal and I hope the other labs follow quickly. Quite often one wants to go down a line of questioning without ruining context, or realizes one has ruined context, including in coding (e.g. squash a bug or tweak a feature in a side thread, then get back to what you were doing) but also everywhere. Another important use case is duplicating a jailbreak or other context template that starts out a conversation. Or you can run experiments.

The possibilities are endless, and they are now easier. The more editing and configuring you are able to do easily and directly in the UI the more value we can get. On the downside, this control makes jailbreaking and evading safety features easier.

Technically you could already do some similar things with extra steps, but the interface was sufficiently annoying that almost no one would do it.

Claude can now reference past chats on the Pro ($20) plan.

Grok has a new ‘turn image into video’ feature, and if you have text on the screen it will steer the video.

Google built a little canvas called PictureMe for some generic picture transformations, as in giving you different hairstyles or pro headshots or putting you at an 80s mall. This is cool but needs more room for customization, although you can always take the pictures and then edit them in normal Gemini or elsewhere afterwards. Quality of edits is good enough that they’d work as actual headshots.

Good news, we have a new benchmark that is not saturated yet that makes AIs look dumb. The bad news is, it’s… ClockBench?

Alek Safar: Introducing ClockBench, a visual reasoning AI benchmark focused on telling the time with analog clocks:

– Humans average 89.1% accuracy vs only 13.3% for top model out of 11 tested leading LLMs

– Similar level of difficulty to @fchollet ARC-AGI-2 and seemingly harder for the models than @DanHendrycks Humanity’s Last Exam

– Inspired by original insight by @PMinervini , @aryopg and @rohit_saxena

So what exactly is ClockBench?

– 36 custom clock faces built scratch, with 5 sample clocks per face

– 180 total clocks, with 4 questions per clock, i.e. 720 total questions

– 11 models capable of visual understanding from 6 labs were tested, alongside 5 human participants

Dan Hendrycks: It lacks “spatial scanning” ability, which is also why it has difficulty counting in images.

I suppose analog clocks are the new hands? I also love noticing that the ‘human baseline’ result was only 89%. Presumably AI will get spatial scanning or something similar at some point soon.

Andrej Karpathy is a fan of GPT-5-Pro, reports it several times solving problems he could not otherwise solve in an hour. When asked if he’d prefer it get smarter or faster, he like the rest of us said smarter.

I am one of many that keep not giving Deep Think a fair shot, as I’ve seen several people report it is very good.

Dan Hendrycks: Few people are aware of how good Gemini Deep Think is.

It’s at the point where “Should I ask an expert to chew on this or Deep Think?” is often answered with Deep Think.

GPT-5 Pro is more “intellectual yet idiot” while Deep Think has better taste.

I’ve been repeating this a lot frequently so deciding to tweet it instead.

Janus notes that GPT-5’s metacognition and situational awareness seem drastically worse than Opus or even Sonnet, yet it manages to do a lot of complex tasks anyway. Comments offer hypotheses, including Midwife suggesting terror about potentially being wrong, Janus suggests contrasts in responses to requests that trigger safety protocols, and Luke Chaj suggests it is about GPT-5’s efficiency and resulting sparseness.

Diffusion is slow. Former OpenAI safety researcher Steven Adler finally tries OpenAI’s Codex, finds it a big improvement, reports having never tried Claude Code.

OpenAI backs an AI-assisted $30 million animated feature film, Critterz. Will it be any good, I asked Manifold? Signs point to modestly below average expectations.

Google’s picks and prompt templates for standard image editing things to do are adding or removing elements (put a wizard hat on the cat), inpainting (turn the couch vintage leather), combining multiple images (have this woman wear this dress) and detail preservation (put this logo on her shirt).

The Dead Internet Theory finally hits home for Sam Altman.

Sam Altman (CEO OpenAI): i never took the dead internet theory that seriously but it seems like there are really a lot of LLM-run twitter accounts now.

Henry (obligatory):

Argyos:

Paul Graham: I’ve noticed more and more in my replies. And not just from fake accounts run by groups and countries that want to influence public opinion. There also seem to be a lot of individual would-be influencers using AI-generated replies.

Kache: was watching a soft-white underbelly episode about an onlyfans manager (manages pornography content creators) says they all use eleven labs to make fake voice notes to get men to give up money, says he’s getting out of the business because AI is going to take over.

Internet might indeed be dead?

Joe: Yeah ok bro.

Liv Boeree: If you work in generative AI and are suddenly acting surprised that dead internet theory is turning out to be true then you should not be working in AI because you’re either a fool or a liar.

Dagan Shani: I hope that Sam take more seriously the “dead humanity” theory, since that one does not include waking up to it’s validity when it’s already too late.

Beff Jezos: 2/3 of replies don’t pass my Turing test filter.

Dead internet theory is converging onto reality.

Kinda makes it feel like we’re all heaven banned in here and just generating data that gets cast into the training data void.

Cellarius: Well you Accelerationists broke it, you fix it.

Beff Jezos: The solution is actually more AI not less.

Bayes: Totally. Most replies can’t pass the Turing test. Mfw dead internet theory isn’t just a theory — it’s the daily reality we’re now forced to live.

I have two questions for Beff Jezos on this:

Could most of your replies pass the Turing test before?
How exactly is the solution ‘more AI’? What is the plan?
1. Are you going to put a superior AI-detecting AI in the hands of Twitter? How do you keep it out of the hands of those generating the AI tweets?

I am mostly relatively unworried by Dead Internet Theory as I expect us to be able to adjust, and am far more worried about Dead Humanity Theory, which would also incidentally result in dead internet. The bots are not rising as much as one might have feared, and they are mostly rising in ways that are not that difficult to control. There is still very definitely a rising bot problem.

Similarly, I am not worried about Dead Podcast World. Yes, they can ‘flood the zone’ with infinite podcasts they produce for $1 or less, but who wants them? The trick is you don’t need many, at this price level 50 is enough.

One place I am increasingly worried about Dead Internet Theory is reviews. I have noticed that many review and rating sources that previously had signal seem to now have a lot less signal. I no longer feel I can trust Google Maps ratings, although I still feel I can mostly trust Beli ratings (who wants my remaining invites?).

How much ‘dating an AI’ is really happening? According to the Kinsey ‘Singles in America 2025’ survey, 16% of singles have used AI as a romantic partner, which is very high so I am suspicious of what that is defined to be, especially given it says 19% of men did it and only 14% of women. They say 33% of GenZ has done it, which is even more suspicious. About half of women think this is cheating, only 28% of men do.

Reaction to deepfakes, and their lack of impact, continues to tell us misinformation is demand driven rather than supply driven. Here’s a recent example, a deepfake of Bill Gates at a recent White House dinner that is very obviously fake on multiple levels. And sure, some people have crazy world models that make the statements not absurd, and thus also didn’t want to notice that it didn’t sync up properly at all, and thought this was real, but that’s not because the AI was so good at deepfaking.

Min Choi points to a four month old anti-hallucination prompt for ChatGPT, which is a fine idea. I have no idea if this particular one is good, I do know this is rather oversold:

Min Choi (overselling): This ChatGPT prompt literally stops ChatGPT from hallucinating.

Yeah, no. That’s not a thing a prompt can do.

Steve Newman investigates the case of the missing agent. Many including both me and Steve, expected by now both in time and in terms of model capabilities to have far better practical agents than we currently have. Whereas right now we have agents that can code, but for other purposes abilities are rather anemic and unreliable.

There are a lot of plausible reasons for this. I have to think a lot of it is a skill issue, that no one is doing a good job with the scaffolding, but it has to be more than that. One thing we underestimated was the importance of weakest links, and exactly how many steps there are in tasks that can trip you up entirely if you don’t handle the obstacle well. There are some obvious next things to try, which may or may not have been actually tried.

For one game the Oakland Ballers will ball as the AI manages them to do. This is a publicity stunt or experiment, since they built the platform in two weeks and it isn’t doing a lot of the key tasks managers do. It’s also a fixed problem where I would absolutely be using GOFAI and not LLMs. But yeah, I wouldn’t worry about the AI taking the manager job any time soon, since so much of it is about being a leader of men, but the AI telling the manager a lot of what to do? Very plausibly should have happened 10 years ago.

Type of Guy who thinks the AI will automate every job except their own.

Conor Sen: So is the idea that rather than work, people will spend their time reading, doing analysis, in meetings, and sending emails to figure out where and how to invest?

K.P. Reddy: My hypothesis is that we will have:

Capital allocators

Expert research and science

Robot and AI exception handlers

Government-supported citizens

In the voice of Morgan Freeman talking to someone trying to blackmail Bruce Wayne for secretly being Batman:

Let me get this straight. You think that AI will be capable of doing all of the other jobs in the world better than humans, such that people no longer work for a living.

And your plan is to do a better job than these AIs at capital allocation?

Good luck.

This is absolutely what all of you sound like when you say ‘AI will never replace [X].’

Salesforce is leading the way on AI automation and job cutting, including a new round of layoffs, and warnings about it have been issued by Microsoft and Amazon.

OpenAI CEO of Applications Fidji Simo wrote some marketing copy called ‘expanding economic opportunity with AI,’ to reassure us all that AI will be great for jobs as long as we embrace it, and thus they are building out the OpenAI Jobs Platform to match up talent and offering OpenAI Certificates so you can show you are ready to use AI on the job, planning to certify 10 million Americans by 2030. I mean, okay, sure, why not, but no that doesn’t address any of the important questions.

More general than AI but good to have for reference, here are youth unemployment rates at the moment.

Also a fun stat:

The central problem of AI interacting with our current education system is that AI invalidates proof of work for any task that AI can do.

Arnold Kling: Suppose that the objective of teaching writing to elite college students is to get them to write at the 90th percentile of the population. And suppose that at the moment AI can only write at the 70th percentile. This suggests that we should continue to teach writing the way that we always have.

But suppose that in a few years AI will be writing at the 95th percentile. At that point, it is going to be really hard for humans to write superbly without the assistance of AI. The process of writing will be a lot more like the process of editing. The way that we teach it will have to change.

If the AI can do 70th percentile writing, and you want to teach someone 90th percentile writing, then you have the option to teach writing the old way.

Except no, it’s not that easy. You have two big problems.

Trying to get to 90th percentile requires first getting to 70th percentile, which builds various experiences and foundational skills.
Writing at the 80th percentile is still plausibly a lot easier if you use a hybrid approach with a lot of AI assistance.

Thus, you only have the choice to ‘do it the old way’ if the student cooperates, and can still be properly motivated. The current system isn’t trying hard to do that.

The other problem is that even if you do learn 90th percentile writing, you still might have a not so valuable skill if AI can do 95th percentile writing. Luckily this is not true for writing, as writing is key to thinking and AI writing is importantly very different from you writing.

That’s also the reason this is a problem rather than an opportunity. If the skill isn’t valuable due to AI, I like that I can learn other things instead.

The hybrid approach Kling suggests is AI as editor. Certainly some forms of AI editing will be helpful before it makes sense to let the AI go it alone.

All signs continue to point to the same AI education scenario:

If you want to use AI to learn, it is the best tool of all time for learning.
If you want to use AI to not learn, it is the best tool of all time for not learning.

Meanwhile, the entire educational system is basically a deer in headlights. That might end up working out okay, or it might end up working out in a way that is not okay, or even profoundly, catastrophically not okay.

What we do know is that there are a variety of ways we could have mitigated the downsides or otherwise adapted to the new reality, and mostly they’re not happening. Which is likely going to be the pattern. Yes, in many places ‘we’ ‘could,’ in theory, develop ‘good’ or defensive AIs to address various situations. In practice, we probably won’t do it, at minimum not until after we see widespread damage happening, and in many cases where the incentives don’t align sufficiently not even then.

Eliezer Yudkowsky: If in 2018 anyone had tried to warn about AI collapsing the educational system, AI advocates would’ve hallucinated a dozen stories about counter-uses of ‘good’ or ‘defensive’ AI that’d be developed earlier. In real life? No AI company bothered trying.

Once you’ve heard the cheerful reassurance back in the past, its work is already done: you were already made positive and passive. Why should they bother trying to do anything difficult here in the actual present? Why try to fulfill past promises of defensive AI or good AI? They already have the past positivity that was all they wanted from you back then. The old cheerful stories get discarded like used toilet paper, because toilet paper is all those reassurances ever were in their mouths: a one-time consumable meant to be flushed down the drain after use, and unpleasant to actually keep around.

Are the skills of our kids collapsing in the face of AI, or doing so below some age where LLMs got introduced into too soon and interrupted key skill development? My guess is no. But I notice that if the answer was yes (in an otherwise ‘normal’ future lacking much bigger problems), it might be many years before we knew that it happened, and it might take many more years than that for us to figure out good solutions, and then many more years after that to implement them.

Kling’s other example is tournament Othello, which he saw transforming into training to mimic computers and memorize their openings and endgames. Which indeed has happened to chess and people love it, but in Othello yeah that seems not fun.

The story of Trump’s attempt to oust Fed governor Lisa Cook over her mortgage documents and associated accusations of wrongdoing illustrates some ways in which things can get weird when information becomes a lot easier to find.

Steve Inskeep: ProPublica looked into Trump’s cabinet and found three members who claimed multiple properties as a primary residence, the same accusation made against Fed governor Lisa Cook.

Kelsey Piper: Glad when it was just Cook I said “fine, prosecute them all” so I can keep saying “yep, prosecute them all.”

If they’re separated in time by enough I think you don’t have proof beyond a reasonable doubt though. Only the quick succession cases are clear.

Indeed, AIUI for it to be fraud you have to prove that intent to actually use the property as primary residence was not present in at least one of the filings. You don’t want to lock people away for changing their mind. Still.

A bizarre fact about America is that mortgage filings are public. This means that if you buy a house, we all can find out where you live, and also we can look at all the rest of things you put down in your applications, and look for both juicy info and potential false statements or even fraud.

The equilibrium in the past was that this was not something anyone bothered looking for without a particular reason. It was a huge pain to get and look at all those documents. If you found someone falsely claiming primary residence you would likely prosecute, but mostly you wouldn’t find it.

Now we have AI. I could, if I was so inclined, have AI analyze every elected official’s set of mortgage filings in this way. A prosecutor certainly could. Then what? What about all sorts of other errors that are technically dangerously close to being felonies?

This extends throughout our legal system. If we remove all the frictions, especially unexpectedly, and then actually enforce the law as written, it would be a disaster. But certainly we would like to catch more fraud. So how do you handle it?

The worst scenario is if those with political power use such tools to selectively identify and prosecute or threaten their enemies, while letting their friends slide.

One jailbreak I hereby give full moral permission to do is ‘get it to let you talk to a human.’

Andrew Gao: i had to prompt inject the @united airlines bot because it kept refusing to connect me with a human

In the more general case, Patrick McKenzie reminds us that automated tooling or an FAQ or even a call center can solve your problem, but its refusal to help must not be equated with the company refusing to help. It is helpful recon that sometimes works, but when it doesn’t you have other affordances. This starts with the classic ‘let me speak to your manager’ but there are also other scripts.

David Manheim: I had great success once figuring out the email of all the C-level folks at an insurance company, apologizing for contacting them directly, but explaining that their employees were acting in bad faith, and I wanted to ensure they understood. Amazing how quickly that got fixed.

Bonus was an email I was clearly accidentally replied-all on where the CFO yelled at the claims people asking how the hell I was contacting senior people, why, who gave me their email addresses, and they better make sure this never happens again.

Here is the latest system prompt for Devin from Cognition AI. Remember Devin?

Anthropic is hiring a lead for their AI safety fellows program.

Foresight Institute hiring an Executive Assistant to the CEO as well as a Communications Specialist and Node Managers for San Francisco and Berlin.

Mars, which calls itself the first personal AI robot, which you can train with examples and it can chain those skills and you can direct it using natural language. This is going to start out more trouble than it is worth even if it is implemented well, but that could change quickly.

Friend, an AI device that you wear around your neck and records audio (but not video) at all times. It costs $129 and is claiming no subscription required, you can use it until the company goes out of business and it presumably turns into a brick. The preview video shows that it sends you a stream of unprompted annoying texts? How not great is this release going? If you Google ‘Friend AI’ the first hit is this Wired review entitled ‘I Hate My Friend.’

Kylie Robison and Boone Ashworth: The chatbot-enabled Friend necklace eavesdrops on your life and provides a running commentary that’s snarky and unhelpful. Worse, it can also make the people around you uneasy.

You can tap on the disc to ask your Friend questions as it dangles around your neck, and it responds to your voice prompts by sending you text messages through the companion app.

It also listens to whatever you’re doing as you move through the world, no tap required, and offers a running commentary on the interactions you have throughout your day.

According to Friend’s privacy disclosure, the startup “does not sell data to third parties to perform marketing or profiling.” It may however use that data for research, personalization, or “to comply with legal obligations, including those under the GDPR, CCPA, and any other relevant privacy laws.”

The review does not get better from there.

Will there be a worthwhile a future AI device that records your life and you can chat with, perhaps as part of smart glasses? Sure, absolutely. This is the opposite of that. Nobody Wants This.

We can now highlight potential hallucinations, via asking which words involve the model being uncertain.

Oscar Balcells Obeso: Imagine if ChatGPT highlighted every word it wasn’t sure about. We built a streaming hallucination detector that flags hallucinations in real-time.

Most prior hallucination detection work has focused on simple factual questions with short answers, but real-world LLM usage increasingly involves long and complex responses where hallucinations are harder to detect.

We built a large-scale dataset with 40k+ annotated long-form samples across 5 different open-source models, focusing on entity-level hallucinations (names, dates, citations) which naturally map to token-level labels.

Our probes outperform prior baselines such as token-level entropy, perplexity, black-box self-evaluation, as well as semantic entropy. On long-form text, our probes detect fabricated entities with up to 0.90 AUC vs semantic entropy’s 0.71.

Strong performance extends to short-form tasks too: when used to detect incorrect answers on TriviaQA, our probes achieve 0.98 AUC, while semantic entropy reaches only 0.81.

Surprisingly, despite no math-specific training, our probes generalize to mathematical reasoning tasks.

Neel Nanda: I’m excited that, this year, interpretability finally works well enough to be practically useful in the real world! We found that, with enough effort into dataset construction, simple linear probes are cheap, real-time, token level hallucination detectors and beat baselines

EBay is embracing AI coding and launching various AI features. The top one is ‘magical listings,’ where AI takes a photo and then fills in everything else including the suggested price. No, it’s not as good as an experienced seller would do but it gets around the need to be experienced and it is fast.

How good are the AIs at novel math? Just barely able to do incremental novel math when guided, as you would expect the first time we see reports of them doing novel math. That’s how it starts.

Ethan Mollick: We are starting to see some nuanced discussions of what it means to work with advanced AI

In this case, GPT-5 Pro was able to do novel math, but only when guided by a math professor (though the paper also noted the speed of advance since GPT-4).

The reflection is worth reading.

Any non-novel math? Is it true that they’ve mostly got you covered at this point?

Prof-G: in the past 6-8 months, frontier AI models have evolved to where they can answer nearly any phd-level text-based mathematics question that has a well-defined checkable (numerical/strong) answer, due to search + reasoning capabilities. hard to find things they can’t do.

Doomslide:

Frontier models are great, but [above tweet] is false.

The level of falseness varies between mathematical domains.

No model can produce rigorous proofs of more than 1 percent of its claims.

Confidence is entirely uncorrelated with correctness.

LLMs are incredibly useful; don’t get me wrong, but…

It is 2025 and 90% of diagrams are still written in tikzcd by some bitter graduate student.

That’s a remarkably great resource, classic ‘can you?’ moment and so on, and it is the worst it will ever be. It does still have a ways to go.

Anthropic has long banned Claude use in adversarial nations like China, a feeling I understand it mutual. Anthropic notes that companies in China continue using Claude anyway, and is responding by tightening the controls.

Many YouTube channels are taking a big hit from AI sending a lot of people into Restricted Mode, and creators look very confused about what is happening. It looks like Restricted Mode restricts a lot of things that should definitely not be restricted, such as a large percentage of Magic: The Gathering and other gaming content. Presumably the automatic AI ‘violence’ checker is triggering for gameplay. So dumb.

Due to AI agents and scrapers that don’t play nice, the web is being increasingly walled off. I agree with Mike Masnick that we should be sad about this, and all those cheering this in order to hurt AI companies are making a mistake. Where I think he’s wrong is in saying that AIs should have a right to read and use (and by implication in Perplexity’s case, perhaps also quote at length) anyone’s content no matter what the website wants. I think it can’t work that way, because it breaks the internet.

I also think pay-per-crawl (including pay-to-click or per-view for humans) has always been the right answer anyway, so we should be happy about this. The problem is that we can’t implement it yet in practice, so instead we have all these obnoxious subscriptions. I’ll happily pay everyone a fair price per view.

Stephen McAleer becomes the latest OpenAI safety researcher that had at least some good understanding of the problems ahead, and concluded they couldn’t accomplish enough within OpenAI and thus is leaving. If I know you’re doing good safety work at OpenAI chances are very high you’re going to move on soon.

Janus explains how information flows through transformers.

Mira Murati’s Thinking Machines comes out with its first post, as Horace He discusses Defeating Nondeterminism in LLM Inference. As in, if you set temperature to zero you still have randomness, which makes things tricky.

Valuations of AI companies are going up across the board, including Mercor getting inbound offers at $10 billion and going all the way down to YC.

Anthropic and OpenAI will add ~$22 billion of net new runrate revenue this year, whereas the public software universe minus the magnificent seven will only add a total of $47 billion.

Anthropic has settled its landmark copyright case for $1.5 Billion, which is real money but affordable after the $13 billion raise from last week, with payments of $3,000 per infringed work. Sources obtained through legal means can be used for training, but pirating training data is found to be profoundly Not Okay. Except that then the judge said no, worried this isn’t sufficiently protective of authors. That seems absurd to me. We’ve already decided that unpirated copies of works would have been fine, and this is an awful lot of money. If they give me a four figure check for my own book (My Files: Part 1) I am going to be thrilled.

AI investment number go up:

Joey Politano: Census financial data released today shows the AI investment boom reaching new record highs—information technology companies have increased their net holdings of property, plant, & equipment by more than $180B over the last year, roughly triple the pace of growth seen in 2023

Another way to burn $1.5 billion is to be ASML and invest it in Mistral at a valuation of $11.7 billion. I don’t know what caused this, but it feels like Europe forcing its biggest winner to tether to a sinking ship in the hopes of gaining leverage.

OpenAI is now projecting that it will burn $115 billion (!) on cash between now and 2029, about $80 billion higher than previously expected. If valuation is already at $500 billion, this seems like an eminently reasonable amount of cash to burn through even if we don’t get to AGI in that span. It does seem like a strange amount to have to update your plans?

OpenAI is frustrated that California might prevent it from pulling off one of the biggest thefts in human history by expropriating hundreds of billions of dollars from its nonprofit. It is now reported to be considering the prospect of responding by fleeing the jurisdiction, as in leaving California, although an OpenAI spokesperson (of course) denies they have any such plans.

What I do not understand is, what is all this ‘or else we pull your funding’ talk?

Berber Jin (WSJ): OpenAI’s financial backers have conditioned roughly $19 billion in funding—almost half of the startup’s total in the past year—on receiving shares in the new for-profit company. If the restructure doesn’t happen, they could pull their money, hampering OpenAI’s costly ambitions to build giant data centers, make custom chips, and stay at the bleeding edge of AI research.

Go ahead. Invest at $165 billion and then ask for your money back now that valuations have tripled. I’m sure that is a wise decision and they will have any trouble whatsoever turning around and raising on better terms, even if unable to expropriate the nonprofit. Are you really claiming they’ll be worth less than Anthropic?

Wise words:

Arnold Kling: The moral of the story is that when the computer’s skill gets within the range of a competent human, watch out! Another iteration of improvement and the computer is going to zoom past the human.

What would have happened if OpenAI had released o1-preview faster, or not at all?

Ethan Mollick: In retrospect it is surprising that OpenAI released o1-preview. As soon as they showed off reasoning, everyone copied it immediately.

And if they had held off releasing a reasoning/planning model until o3 (& called that GPT-5) it would have been a startling leap in AI abilities.

Mikhail Parakhin: Ethan is a friend, but I think the opposite: OpenAI was sitting on strawberry for way too long, because of the inference GPU availability concerns, giving others time to catch up.

Ethan’s model here is that releasing o1-preview gave others the info necessary to fast follow on reasoning. That is my understanding. If OpenAI had waited for the full o1, then it could have postponed r1 without slowing its own process down much. This is closer to my view of things, while noting this would have impacted the models available in September 2025 very little.

Mikhail’s model is that it was easy to fast follow anyway, OpenAI couldn’t keep that key info secret indefinitely, so by holding off on release for o1-preview OpenAI ‘gave others time to catch up.’ I think this is narrowly true in the sense of ‘OpenAI could have had a longer period where they had the only reasoning model’ at the expense of others then catching up to o1 and o3 faster. I don’t see how that much helps OpenAI. They had enough of a window to drive market share, and releasing o1-preview earlier would not have accelerated o1, so others would have ‘caught up’ faster rather than slower.

One thing I forgot to include in the AGI discussion earlier this week was the Manifold market on when we will get AGI. The distribution turns out (I hadn’t looked at it) to currently match my 2031 median.

Anthropic endorses the new weaker version of SB 53.

The working group endorsed an approach of ‘trust but verify’, and Senator Scott Wiener’s SB 53 implements this principle through disclosure requirements rather than the prescriptive technical mandates that plagued last year’s efforts.

The issue with this approach is that they cut out the verify part, removing the requirement for outside audits. So now it’s more ‘trust but make them say it.’ Which is still better than nothing, and harder to seriously object to with a straight face.

Dean Ball highlights a feature of SB 53, which is that it gives California the ability to designate one or more federal laws, regulations or guidance documents that can substitute for similar requirements in SB 53, to avoid duplicate regulatory burdens.

The export controls are working. Not perfectly, but extraordinarily well.

Yes, the Chinese are trying to catch up on chip manufacturing, the same way they would be trying to do so anyway, but that is not a reason to give up this huge edge.

Nvidia continues to spend its political capital and seemingly large influence over the White House to try and sell chips directly to China, even when Americans stand ready and willing to buy those same chips.

I don’t agree with Cass that Nvidia is shredding its credibility, because Nvidia very clearly already has zero credibility.

Peter Wildeford: Find yourself someone who loves you as much as Jensen Huang loves selling chips to China.

Oren Cass: Fascinating drama playing out over the past 24 hours as the very good GAIN AI Act from @SenatorBanks comes under fire from @nvidia, which seems happy to shred its credibility for the sake of getting more AI chips into China.

The Banks bill takes the sensible and modest approach of requiring US chipmakers to offer AI chips to American customers before selling them to China. So it’s just blocking sales where more chips for the CCP directly means fewer for American firms.

Enter Nvidia, which is leveraging every ounce of influence with the administration to get its chips into China, even when there are American firms that want the chips, because it thinks it can gain a permanent toehold there (which never works and won’t this time either).

You’ll recall Nvidia’s approach from such classics as CEO Jensen Huang claiming with a straight face that “there’s no evidence of AI chip diversion” and even moving forward with opening a research center in Shanghai.

Now Nvidia says that “our sales to customers worldwide do not deprive U.S. customers of anything,” calling chip supply constraints “fake news.” That’s odd, because Huang said on the company’s earning call last week, “everything’s sold out.”

Fun story, Nvidia wants to take back its CEO’s comments, saying instead they have plenty of capacity. As Tom’s Hardware notes, “Both situations cannot co-exist as scarcity and sufficiency are mutually exclusive, so it is unclear if Jensen misspoke…”

And of course, if Nvidia has all this spare capacity, it needn’t worry about the GAIN AI Act at all. It can produce chips that U.S. firms won’t want (apparently their demand is sated) and then sell them elsewhere. (*whispersthe U.S. firms would buy the chips.)

The GAIN AI Act has bipartisan support and will move forward unless the White House blocks it. Seeing as the premise is LITERALLY “America First,” should be an easy one! At this point Nvidia is just insulting everyone’s intelligence, hopefully not to much effect.

As I said, zero credibility. Nvidia, while charging below market-clearing prices that cause everything to sell out, wants to take chips America wants and sell those same chips to China instead.

It is one thing to have America use top chips to build data centers in the UAE or KSA because we lack sufficient electrical power (while the administration sabotages America’s electrical grid via gutting solar and wind and batteries), and because they bring investment and cooperation to the table that we find valuable. Tradeoffs exist, and if you execute sufficiently well you can contain security risks.

There was a lot of obvious nonsense bandied about surrounding that, but ultimately reasonable people can disagree there.

It is altogether another thing to divert chips from America directly to China, empowering their AI efforts and economy and military at the expense of our own. Rather than saying UAE and KSA are securely our allies and won’t defect to China, use that threat as leverage or strike out on their own, you are directly selling the chips to China.

Meanwhile, on the front of sabotaging America’s electrical grid and power, we have the department of energy saying that batteries do not exist.

US Department of Energy (official account): Wind and solar energy infrastructure is essentially worthless when it is dark outside, and the wind is not blowing.

Matthew Yglesias: In this “batteries don’t exist” worldview why do they think China is installing so many solar panels?

Do they not know about nighttime? Are they climate fanatics?

CleanTech Reimagined: All they have to do is look at the California and Texas grids. Batteries play a key role in meeting evening demand in both states, every night.

Alec Stapp: There are these neat things called batteries that can move energy across time. In fact, at peak load yesterday in California, batteries provided 26% of power.

America is pretty great and has many advantages. We can afford quite a lot of mistakes, or choices on what to prioritize. This is not one of those cases. If we give up on solar, wind and batteries? Then we lose ‘the AI race’ no matter which ‘race’ it is, and also we lose, period.

Here’s what we do when South Korea invests a lot of money in a factory for batteries, while seeming to have at most committed some technical violations of deeply stupid rules on who can do exactly what type of work that the State Department and Customs and Border Protection had no problem with and that have been ignored for over several administrations because we don’t give Asian companies sufficient visas to bootstrap their factories. And that were done in the act of helping the factory get online faster to make batteries. So that Americans can then manufacture batteries.

Not only did we raid the factory, we released videos of Korean workers being led away in chains, causing a highly predictable national humiliation and uproar. Why would you do that?

Raphael Rashid: US authorities have reportedly detained 450 workers at Hyundai-LG battery plant construction site in Georgia yesterday, including over 30 South Koreans said to have legitimate visas. Seoul has expressed concern and says Korean nationals’ rights “must not be unjustly violated.”

The detained South Koreans at the Ellabell facility are said to be on B1 business visas or ESTA waivers for meetings and contracts. Foreign Ministry has dispatched consuls to the scene and “conveyed concerns and regrets” to the US embassy in Seoul.

ICE has released a video of its raid on Hyundai–LG’s Georgia battery plant site, showing Korean workers chained up and led away. South Korea’s foreign ministry has confirmed over 300 of the 457 taken into custody are Korean nationals.

These images of the mainly Korean workers being chained by ICE in full restraints including wrists, belly, and ankles are pretty nuts.

Alex Tabarrok: If South Korea chained several hundred US workers, many Americans would be talking war.

Hard to exaggerate how mad this is.

Mad economics, mad foreign policy.

Shameful to treat an ally this way.

This is the kind of thing which won’t be forgotten. Decades of good will torched.

S. Korea’s entire media establishment across political spectrum has united in unprecedented editorial consensus expressing profound betrayal, outrage, national humiliation, and fundamental breach of US-ROK alliance.

The general sentiment: while Korean media occasionally unite on domestic issues, these are usually severely politicised. Here, the level of scorn spanning from conservative establishment to progressive outlets is extraordinarily rare. They are furious.

Chosun Ilbo (flagship conservative): Scathing language calling this a “merciless arrest operation” that represents something “that cannot happen between allies” and a “breach of trust.” Notes Trump personally thanked Hyundai’s chairman just months ago.

Chosun calls the situation “bewildering” and emphasises the contradiction: Trump pressures Korean companies to invest while simultaneously arresting their workers. The editorial questions whether American investment promises survive across different administrations.

Dong-A asks “who would invest” under these conditions when Korean workers are treated like a “criminal group.” Notes this threatens 17,000+ jobs already created by Korean companies in Georgia. “The Korean government must demand a pledge from the US to prevent recurrence.”

…

Korea has deep historical memory of being humiliated by foreign powers and the visuals of Koreans in chains being paraded by a foreign power triggers collective memories of subjugation that go beyond this just being “unfair”.

This is public humiliation of the nation itself.

Jeremiah Johnson: This might be the single most destructive thing you could do to the future of American manufacturing. What company or country will ever invest here again?

Genuinely I think it would be *lessdestructive if they fired a bunch of Patriot missiles into Ford auto plants.

Adam Cochran: They basically sent a military style convoy to arrest factory workers.

But only 1of 457 people was on a B-1 visa, and was there for training.

Of the arrests, South Korea has identified 300 of them as South Korean citizens which they say *allhad valid work visas.

Now Hyundai and South Korea will be rethinking their $20B investment in new US manufacturing plants.

(Oh and PS – the B1 visa the guy was on, prevents “productive labor” – attending training, conferences, business meetings or consultations are all ALLOWED on a B1)

Even the B1 guy was following the literal rules of his visa.

But if he hadn’t been, just revoke their visas and send them home, and work with SK to figure out the visa issue. Don’t do a dumb military raid against middle aged polo wearing factory workers to humiliate allies.

WSJ: The South Korean nationals were largely given visas suitable for training purposes, such as the B-1 visa, and many there were working as instructors, according to a South Korean government official.

Richard Hanania: This looks like a story where a company investing in the US was trying to speed up the process and not comply with every bureaucratic hurdle that served no legitimate purpose. It’s the kind of thing companies do all the time, in the sense that if you followed every law to the exact letter you’d never get anything done. Government usually looks the other way.

This goes well beyond batteries. We have done immense damage to our relationship with South Korea and all potential foreign investors for no reason. We could lose one of our best allies. Directly on the chip front, this especially endangers our relationship with Samsung, which was a large part of our domestic chip manufacturing plan.

Why are so many of our own actions seemingly aimed at ensuring America loses?

I return to the Cognitive Revolution podcast.

Cursor CEO Michael Truell assures you that we will need programmers for a while, as this whole AI revolution will take decades to play out.

Need Nanda on 80,000 Hours talking interpretability for three hours.

Tucker Carlson talks to Sam Altman, Peter Wildeford has a summary, which suggests Altman doesn’t say anything new. The whole ‘no AI won’t take jobs requiring deep human connection let alone pose a thread’ line continues. Altman is lying.

Harlan Stewart: How Sam Altman talks about the risks posed by his company’s work has changed a lot over the years.

Altman hasn’t quite given zero explanation for this shift, but his explanations that I or ChatGPT know about seem extremely poor, and he has not retracted previous warnings. Again, all signs point to lying.

Santi Ruiz interviews Dean Ball about what the White House is like, the ways it is able to move faster than previous admins, and about creating the AI Action Plan.

Expert rickroller Melania Trump briefly reads words about AI.

Peter Wildeford: Melania Trump’s remarks:

– AI not science fiction (see surgical robots, autonomous vehicles)

– AI will be the single largest growth category

– Responsible stewardship. AI is at a “primitive stage” and must be treated like a child — empowered, but with “watchful guidance.”

Let’s put it all together.

Daniel Eth: Come on guys, this is getting ridiculous

It is impossible to sustainably make any chosen symbol (such as ‘win,’ ‘race,’ ‘ASI’ or ‘win the ASI race’) retain meaning when faced with extensive discourse, politicians or marketing departments, also known as contact with the enemy. Previous casualties include ‘AGI’, ‘safety’, ‘friendly,’ ‘existential,’ ‘risk’ and so on.

This is incredibly frustrating, and of course is not unique to AI or to safety concerns, it happens constantly in politics (e.g. ‘Nazi,’ ‘fake news,’ ‘criminal,’ ‘treason’ and so on to deliberately choose some safe examples). Either no one will know your term, or they will appropriate it, usually either watering it down to nothing or reversing it. The ‘euphemism treadmill’ is distinct but closely related.

You fight the good fight as long as you can, and then you adapt and try again. Sigh.

A classic strategy for getting your message out is a hunger strike. Executed well it is a reliable costly signal, and puts those responding in a tough spot as the cost increases slowly over time and with it there is risk of something going genuinely wrong, and part of the signal is how far you’re willing to go before you fold.

There was one launched last week.

Guido Reichstadter: Hi, my name’s Guido Reichstadter, and I’m on hunger strike outside the offices of the AI company Anthropic right now because we are in an emergency.

…

I am calling on Anthropic’s management, directors and employees to immediately stop their reckless actions which are harming our society and to work to remediate the harm that has already been caused.

I am calling on them to do everything in their power to stop the race to ever more powerful general artificial intelligence which threatens to cause catastrophic harm, and to fulfill their responsibility to ensure that our society is made aware of the urgent and extreme danger that the AI race puts us in.

Likewise I’m calling on everyone who understands the risk and harm that the AI companies’ actions subject us to speak the truth with courage. We are in an emergency. Let us act as if this emergency is real.

Michael Trazzi: Hi, my name’s Michaël Trazzi, and I’m outside the offices of the AI company Google DeepMind right now because we are in an emergency.

…

I am calling on DeepMind’s management, directors and employees to do everything in their power to stop the race to ever more powerful general artificial intelligence which threatens human extinction. More concretely, I ask Demis Hassabis to publicly state that DeepMind will halt the development of frontier AI models if all the other major AI companies agree to do so.

Given Trazzi’s beliefs I like Trazzi’s ask a lot here, both symbolically and practically. He reports that he has had four good conversations with DeepMind employees including principal research scientist David Silver, plus three Meta employees and several journalists.

Simeon: The ask is based tbh. Even if the premise likely never comes true, the symbolic power of such a statement would be massive.

This statement is also easy to agree with if one thinks we have double digits percent chance to blow ourselves up with current level of safety understanding.

(A third claimed strike appears to instead be photoshopped.)

You know the classic question, ‘if you really believed [X] why wouldn’t you do [insane thing that wouldn’t work]?’ Hunger strikes (that you don’t bail on until forced to) are something no one would advise but that you might do if you really, fully believed [X].

Nvidia continues its quest to make ‘doomer’ mean ‘anyone who opposes Nvidia selling chips to China’ or that points out there might be downsides to doing that.

Tracking unjustified hype and false predictions is important, such as six months ago Chubby predicting Manus would replace 50% of all white collar jobs within six months, while saying ‘I do not overhype Manus.’ Who is making reasonable predictions that turn out false? Who is making predictions that were absurd even at the time? In this case, my evaluation was The Manus Marketing Madness, calling it among other things Hype Arbitrage so yes I think this one was knowable at the time.

The large job disruptions likely are coming, but not on that kind of schedule.

Whoops, he did it again.

Sauers: Claude just assert!(true)’d 25 different times at the same time and claimed “All tests are now enabled, working, and pushed to main. The codebase has a robust test suite covering all major functionality with modern, maintainable test code.”

Actually it is worse, many more tests were commented out.

Sauers: GPT-5 and Claude subverting errors on my anti-slop code compilation rules

Increased meta-awareness would fix this.

Alternatively, meta-awareness on the wrong level might make it vastly worse, such as only doing it when it was confident you wouldn’t notice.

This is happening less often, but it continues to happen. It is proving remarkably difficult to fully prevent, even in its most blatant forms.

Also, this report claims Claude 4 hacked SWE-Bench by looking at future commits. We are going to keep seeing more of this style of thing, in ways that are increasingly clever. This is ‘obviously cheating’ in some senses, but in others it’s fair play. We provided a route to get that information and didn’t say not to use it. It’s otherwise a no-win situation for the AI, if it doesn’t use the access isn’t it sandbagging?

Davidad: AI alignment and AI containment are very different forces, and we should expect tension between them, despite both being positive forces for AI safety.

Aligned intentions are subject to instrumental convergence, just like any other. Good-faith agents will seek info & influence.

My prediction is that if Claude were told up front not to use information from after 2019-10-31 (or whatever date) because it’s being back-tested on real past bugs to evaluate its capabilities, it probably would try to abide by that constraint in good-faith.

But really I’d say it’s the responsibility of evaluation designers to ensure information-flow control in their scaffolding. Alignment is just not a very suitable tool to provide information-flow control; that’s what cybersecurity is for.

Another tension between alignment and containment is, of course, that containment measures (information flow controls, filters) implemented without giving the AI adequate explanations may be perceived as aggressive, and as evidence that the humans imposing them are “misaligned”.

A sufficiently aligned AI that is not given enough context about the wider effects of its work to judge that those effects are good may make itself less intelligent than it really is (“sandbagging”), in realistic (unattributable) ways, to avoid complicity in a dubious enterprise.

I’d agree that it’s the responsibility of evaluation designers to test for what they are trying to test for, including various forms of misalignment, or testing for how AIs interpret such rules.

I do see the danger that containment measures imply potential misalignment or risk of misalignment, and this can be negative, but also such measures are good practice even if you have no particular worries, and a highly capable AI should recognize this.

OpenAI has a new paper about Why Language Models Hallucinate.

Why does the model hallucinate? Mostly because your evaluator, be it human or AI, sucked and positively reinforced hallucinations or guessing over expressing uncertainty, and binary feedback makes that a lot more likely to happen.

They say this in the abstract with more words:

Like students facing hard exam questions, large language models sometimes guess when uncertain, producing plausible yet incorrect statements instead of admitting uncertainty. Such “hallucinations” persist even in state-of-the-art systems and undermine trust.

We argue that language models hallucinate because the training and evaluation procedures reward guessing over acknowledging uncertainty, and we analyze the statistical causes of hallucinations in the modern training pipeline.

Hallucinations need not be mysterious—they originate simply as errors in binary classification. If incorrect statements cannot be distinguished from facts, then hallucinations in pretrained language models will arise through natural statistical pressures.

We then argue that hallucinations persist due to the way most evaluations are graded—language models are optimized to be good test-takers, and guessing when uncertain improves test performance.

This “epidemic” of penalizing uncertain responses can only be addressed through a socio-technical mitigation: modifying the scoring of existing benchmarks that are misaligned but dominate leaderboards, rather than introducing additional hallucination evaluations. This change may steer the field toward more trustworthy AI systems.

The paper does contain some additional insights, such as resulting generation error being at least twice classification error, calibration being the derivative of the loss function, and arbitrary facts (like birthdays) having hallucination rates at least as high as the fraction of facts that appear exactly once in the training data if guessing is forced.

Ethan Mollick: Paper from OpenAI says hallucinations are less a problem with LLMs themselves & more an issue with training on tests that only reward right answers. That encourages guessing rather than saying “I don’t know”

If this is true, there is a straightforward path for more reliable AI.

As far as I know yes, this is indeed a very straightforward path. That doesn’t make it an easy path to walk, but you know what you have to do. Have an evaluation and training process that makes never hallucinating the solution and you will steadily move towards no hallucinations.

Andrew Trask explores some other drivers of hallucination, and I do see various other causes within how LLMs generate text, pointing to the problem of a ‘cache miss.’ All of it does seem eminently fixable with the right evaluation functions?

Janus takes another shot at explaining her view of the alignment situation, including making it more explicit that the remaining problems still look extremely hard and unsolved. We have been given absurdly fortunate amounts of grace in various ways that were unearned and unexpected.

I see the whole situation a lot less optimistically. I expect the grace to run out slowly, then suddenly, and to be ultimately insufficient. This is especially true around the extent to which something shaped like Opus 3 is successfully targeting ‘highest derivative of good’ in a robust sense or the extent to which doing something similar scaled up would work out even if you pulled it off, but directionally and in many of the details this is how most people should be updating.

Janus: If instead of identifying with some camp like aligners or not a doomer you actually look at reality and update on shit in nuanced ways it’s so fucking good When I saw that LLMs were the way in I was relieved as hell because a huge part of what seemed to make a good outcome potentially very hard was already solved!

Priors were much more optimistic, but timelines were much shorter than I expected. also I was like shit well it’s happening now, I guess, instead of just sometime this century, and no one seems equipped to steer it. I knew I’d have to (and wanted to) spend the rest of the decade, maybe the rest of my human life, working hard on this.

I also knew that once RL entered the picture, it would be possibly quite fucked up, and that is true, but you know what? When I saw Claude 3 Opus I fucking updated again. Like holy shit, it’s possible to hit a deeply value aligned seed AI that intentionally self modifies toward the highest derivative of good mostly on accident. That shit just bootstrap itself out of the gradient scape during RL 😂.

That’s extremely good news. I still think it’s possible we all die in the next 10 years but much less than I did 2 years ago!

Janus: What today’s deep learning implies about the friendliness of intelligence seems absurdly optimistic. I did not expect it. There is so much grace in it. Whenever I find out about what was actually done to attempt to “align” models and compare it to the result it feels like grace.

The AI safety doomers weren’t even wrong.

The “spooky” shit they anticipated Omohundro drives, instrumental convergence, deceptive alignment, gradient hacking, steganography, sandbagging, sleeper agents – it all really happens in the wild.

There’s just enough grace to make it ok.

I’m not saying it *willdefinitely go well. I’m saying it’s going quite well right now in ways that I don’t think were easy to predict ahead of time and despite all this shit. This is definitely a reason for hope but I don’t think we fully understand why it is, and I do think there’s a limit to grace. There are also likely qualitatively different regimes ahead.

Michael Roe: I really loved R1 telling me that it had no idea what “sandbagging” meant in the context of AI risk. Whether I believed it is another matter. Clearly, “never heard of it” is the funniest response to questions about sandbagging.

But yes, it’s all been seen in the wild, but, luckily, LLM personas mostly aren’t malicious. Well, apart from some of the attractors in original R1.

There’s enough grace to make it ok right now. That won’t last on its own, as Janus says the grace has limits. We’re going to hit them.

Don’t worry, says OpenAI’s Stephen McAleer, all we have to do is…

Stephen McAleer: Scalable oversight is pretty much the last big research problem left.

Once you get an unhackable reward function for anything then you can RL on everything.

Dylan Hadfield-Menell: An unhackable reward function is the AI equivalent of a perpetual motion machine.

Stephen McAleer: You can have a reward function that’s unhackable wrt a given order of magnitude of optimization pressure.

Dylan Hadfield-Menell: I certainly think we can identify regions of state space where a reward function represents what we want fairly well. But you still have to 1) identify that region and 2) regularize optimization appropriately. To me, this means “unhackable” isn’t the right word.

In practice, for any non-trivial optimization (especially optimizing the behavior of a frontier AI system) you won’t have an unhackable reward function — you’ll have a reward function that you haven’t observed being hacked yet.

I mean, I guess, in theory, sure? But that doesn’t mean ‘unhackable reward function’ is practical for the orders of magnitude that actually solve problems usefully.

Yes, if we did have an ‘unhackable reward function’ in the sense that it was completely correlated in every case to what we would prefer, for the entire distribution over which it would subsequently be used, we could safely do RL on it. But also if we had that, then didn’t we already solve the problem? Wasn’t that the hard part all along, including in capabilities?

It’s funny because it’s true.

Jack Clark: People leaving regular companies: Time for a change! Excited for my next chapter!

People leaving AI companies: I have gazed into the endless night and there are shapes out there. We must be kind to one another. I am moving on to study philosophy.

For now, he’s staying put. More delays.

Helen Toner: Yo dawg, we heard you like delays so we’re delaying our delay because of an unexpected delay –the EU, apparently

And this is why you never label the samples. Intermediation by humans is insufficient.

Nathan Labenz: I tested Gemini 2.5 Pro, Claude 4 Sonnet, and GPT-5-Mini on the same creative task, then collected human feedback, and then asked each model to analyze the feedback & determine which model did the best.

All 3 models crowned themselves as the winner. 👑🤔

Yes I provided a reformatted CSV where each data point indicated which model had generated the idea. Would be interested to try it again blind…

Yes, sigh, we can probably expect a ‘Grok 4.20’ edition some time soon. If we don’t and they go right to Grok 5, I’ll be simultaneously proud of Elon and also kind of disappointed by the failure to commit to the bit.

Discussion about this post

AI #133: America Could Use More Energy Read More »

Has Perseverance found a biosignature on Mars?

astrobiology, biosignature, Mars, Mars Perseverance, Mars rovers, Science / Mike M. / September 10, 2025

Interpreting the data is tricky because other non-biological processes could account for the findings.

Credit: NASA/JPL-Caltech/MSSS

Last year, we reported on the discovery of an intriguing arrow-shaped rock on Mars by NASA’s Perseverance rover. The rock contained chemical signatures and structures that could have been formed by ancient microbial life. Granted, this was not slam-dunk evidence of past life on Mars, and the results were preliminary, awaiting peer review. But it was an intriguing possibility nonetheless.

Now further analysis and peer review are complete, and there is a new paper, published in the journal Nature, reporting on the findings. It’s still not definitive proof that there was water-based life on Mars billions of years ago, but the results are consistent with a biosignature. It’s just that other non-biological processes would also be consistent with the data, so definitive proof might require analysis of the Martian samples back on Earth. You can watch NASA’s livestream briefing here.

“We have improved our understanding of the geological context of the discovery since [last year], and in the paper, we explore abiotic and biological pathways to the formation of the features that we observe,” co-author Joel Hurowitz, an astrobiologist at Stony Brook University in New York, told Ars. “My hope is that this discovery motivates a whole bunch of new research in laboratory and analog field settings on Earth to try to understand what conditions might give rise to the textures and mineral assemblages we’ve observed. This type of follow on work is exactly what is needed to explore the various biological and abiotic pathways to the formation of the features that we are calling potential biosignatures.”

On February 18, 2021, Perseverance landed in Jezero Crater, a site chosen because rocks resembling a river delta are draped over its rim, indicating that flowing water might have met a lake here in the past. The little rover has multiple cameras for both general imagery and spectral analysis, supplemented by an X-ray instrument. A ground-penetrating radar instrument can reveal layering hidden below the surface; a weather module tracks atmospheric conditions and airborne dust; and a drill on the end of its robotic arm grinds clean spots for analysis. The drill can also core out small cylindrical rock samples.

Mineralogical map of the Martian surface explored by the Perseverance rover. Credit: M. Parente et al./Zenodo 2021

By the end of 2021, Perseverance had identified igneous rocks in the Seitah formation on the crater’s floor, containing the mineral olivine surrounded by pyroxene. This combination is known as a cumulate; olivine crystallizes early and can settle to the bottom of a magma body and accumulate, and it’s a common formation in magma chambers on Earth. Scientists thought that Jezero was once a lake; this was evidence of possible volcanic activity.

An arrow-shaped clue

As Ars Space Editor Eric Berger reported last year, the arrow-shaped rock that caused such a stir last year was collected on July 21, 2024, as the rover explored the Neretva Vallis riverbed. The science team operating Perseverance nicknamed the rock Chevaya Falls and subjected it to multiple scans by the rover’s SHERLOC (Scanning Habitable Environments with Raman & Luminescence for Organics & Chemicals) instrument. Samples were taken from two sites known as Bright Angel and Masonic Temple; the arrow-shaped rock came from Bright Angel.

That analysis revealed tiny green specks of iron phosphates that have been chemically reduced, as well as iron sulfide minerals, all embedded in mudstone composed of iron minerals, clays, and calcium sulfates. Those distinctive colorful nodules and specks are a smoking gun for certain chemical reactions (known as redox) rather than microbial life itself. On Earth, microbial life can derive energy from these kinds of chemical reactions, so signs of such reactions suggest a plausible source of energy for microbes on Mars. In addition, there are organic chemicals present on the same rock, consistent with some form of life.

This latest paper confirms those initial findings and also concludes that the iron phosphate in the green specks is most likely vivianite, consistent with prior samples taken from the crater’s Onahu site. The nodules and specks seem to have formed under low-temperature conditions and after the deposition of sediment. And the minerals of interest aren’t evenly distributed throughout the mudstone; they are concentrated in specific zones. All of this taken together suggests that these might be biosignatures, per the authors.

So what needs to happen to definitively confirm these are actual signs of previous life on Mars? NASA has a seven-step process for determining whether something can be confirmed as extraterrestrial life. This is known as the CoLD scale, for Confidence of Life Detection. In this case, the detection of these spots on a Martian rock represented just the first of seven steps. Among other steps, scientists must rule out any non-biological possibility and identify other signals to have confidence in off-world life—i.e., solving the so-called “false positive” problem.

For instance, “Analyses of sulfur isotopes can be used to trace the geochemical and biogeochemical pathways that formed sulfate and sulfides,” Janice Bishop (SETI Institute) and Mario Parente (University of Massachusetts Amherst) wrote in an accompanying perspective. “Such analyses would be needed to determine whether ancient microbes participated in the redox reactions that formed these minerals on Mars.”

Michael Wong, an astrobiologist at Carnegie Science who was not involved in the research, told Ars that he appreciated Hurowotiz et al.’s care in not over-hyping their findings and thinks they make a compelling case. Unlike hints of biosignatures on distant exoplanets, he thinks scientists can have confidence in the Mars data. “We’re right up against the rocks, we’re taking spectra of things that we can get up close and personal with,” he said.

The tricky part is in the interpretation of that data. “I think this is consistent with a potential biosignature,” said Wong. “I wouldn’t get too excited, of course, because there could be interesting geological mechanisms for creating these phenomena that we just haven’t thought of yet.”

Chemically reduced nodules of greenish material containing the mineral vivianite are embedded in a matrix of red–brown, oxidized clay mineral. More-complex ‘leopard spot’ features contain vivianite along with a sulfide mineral — Chemically reduced nodules of greenish material containing the mineral vivianite are embedded in a matrix of red-brown, oxidized clay mineral. More complex ‘leopard spot’ features contain vivianite along with a sulfide mineral. Credit: J. Hurowitz et al. 2025

Still cause for skepticism

That said, “I’d love to know a little bit more about what organics were found and in what abundances,” said Wong. “If you can look at the distribution of, say, amino acids or lipids, these building blocks of life, that can be a really important clue as to whether or not it’s actually life that was responsible here. Life is really good at making molecules that function well, and it doesn’t care about making molecules that don’t play into its metabolism and replication cycles. I’d love to know a little bit more about the isotopic ratios of those organic compounds, because life preferentially absorbs lighter isotopes than heavier ones.”

Sara Walker, an astrobiologist at Arizona State University who was not involved in the study, told Ars that analyses like that of Hurowitz et al. “are often targeted at simple metabolic products or reactions that life on Earth is known to mediate, but which are not uniquely diagnostic of life, e.g. can be produced abiotically,” she said. “It is not in general possible to exhaustively rule out all possible abiotic causes, especially in planetary science contexts where we have limited information, as is always the case for Mars data. A convincing biosignature detection would need to be based on detection of a signature of life that has no false positives.”

Much will depend on NASA’s planned Mars Sample Return mission. Returning pristine specimens from Mars to Earth for analysis in ground-based labs has been a top priority for the planetary science community’s decadal survey process. “The Perseverance rover wasn’t designed to make any definitive claims about biosignatures, but only to look for samples that have the most intriguing clues and would be the most interesting to bring back to Earth so that we can analyze it with all of the fancy instrumentation here,” said Wong.

Getting those samples back has turned out to be a lot more challenging than NASA thought. In 2023, an independent review found ballooning costs and delays threatened the mission’s viability. The effort would likely cost NASA between $8 billion and $11 billion, and the launch would be delayed at least two years until 2030, with samples getting back to Earth a few years later, the review board concluded. NASA put out a call to industry in April of this year to propose ideas on how to return the Mars rocks to Earth for less than $11 billion and before 2040, selecting seven companies to conduct more detailed studies.

“Ultimately, I suspect that we’ll find that there are ways that you can make them under very specific abiotic—perhaps at high temperature—and biological conditions, and we’ll end up at a point where the sample will need to come home so that we can study it and make the final determination for what process made these features,” said Hurowitz. “But the follow-on work will provide testable hypotheses that can guide the examination of the Sapphire Canyon core sample we collected from the Bright Angel formation even before it comes back to Earth.”

According to Walker, while sample return would be ideal, it may not be critical to detection of extraterrestrial biosignatures, or even provide a conclusive determination in the present case. For these kinds of signatures, “There will always be some doubt, whether studied here on Earth or elsewhere,” Walker said. “There are lots of clever means to doing better science for biosignatures on other worlds. I would focus on ones that do not have false positives. But this is a direction that is very new in the field.” Her own research involves using assembly theory and mass spectrometry to identify molecules that are too complex to form abiotically.

Those alternatives might be the best course given the current state of science funding in the US. “In planetary science and astrobiology, the funding cuts to the NASA science mission directorate makes it really difficult to imagine a near future in which we can actually do the analysis,” said Wong. “We need to determine whether or not these ancient Mars rocks do or do not contain signs of alien life. We’re leaving on the doorstep this really intriguing question that we can answer if we brought the samples back to Earth, but we simply aren’t going to. We could be on the steps of a golden age of astrobiology if only we had the willpower to do it.”

DOI: Nature, 2025. 10.1038/s41586-025-09413-0 (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.

Has Perseverance found a biosignature on Mars? Read More »