Kris Guyer – Page 13

Browser extensions with 8 million users collect extended AI conversations

AI, ai bot, Biz & IT, browser extensions, chrome, EDGE, privacy, Security / Kris Guyer / December 17, 2025

Besides ChatGPT, Claude, and Gemini, the extensions harvest all conversations from Copilot, Perplexity, DeepSeek, Grok, and Meta AI. Koi said the full description of the data captured includes:

Every prompt a user sends to the AI
Every response received
Conversation identifiers and timestamps
Session metadata
The specific AI platform and model used

The executor script runs independently from the VPN networking, ad blocking, or other core functionality. That means that even when a user toggles off VPN networking, AI protection, ad blocking, or other functions, the conversation collection continues. The only way to stop the harvesting is to disable the extension in the browser settings or to uninstall it.

Koi said it first discovered the conversation harvesting in Urban VPN Proxy, a VPN routing extension that lists “AI protection” as one of its benefits. The data collection began in early July with the release of version 5.5.0.

“Anyone who used ChatGPT, Claude, Gemini, or the other targeted platforms while Urban VPN was installed after July 9, 2025 should assume those conversations are now on Urban VPN’s servers and have been shared with third parties,” the company said. “Medical questions, financial details, proprietary code, personal dilemmas—all of it, sold for ‘marketing analytics purposes.’”

Following that discovery, the security firm uncovered seven additional extensions with identical AI harvesting functionality. Four of the extensions are available in the Chrome Web Store. The other four are on the Edge add-ons page. Collectively, they have been installed more than 8 million times.

They are:

Chrome Store

Urban VPN Proxy: 6 million users
1ClickVPN Proxy: 600,000 users
Urban Browser Guard: 40,000 users
Urban Ad Blocker: 10,000 users

Edge Add-ons:

Urban VPN Proxy: 1,32 million users
1ClickVPN Proxy: 36,459 users
Urban Browser Guard – 12,624 users
Urban Ad Blocker – 6,476 users

Read the fine print

The extensions come with conflicting messages about how they handle bot conversations, which often contain deeply personal information about users’ physical and mental health, finances, personal relationships, and other sensitive information that could be a gold mine for marketers and data brokers. The Urban VPN Proxy in the Chrome Web Store, for instance, lists “AI protection” as a benefit. It goes on to say:

Browser extensions with 8 million users collect extended AI conversations Read More »

Reporter suggests Half-Life 3 will be a Steam Machine launch title

gaming, half-life, Half-Life 3, hl3, rumor, steam machine, valve / Kris Guyer / December 16, 2025

If you can take your mind way back to the beginning of 2025, you might remember a fresh wave of rumors suggesting that Half-Life 3 was finally reaching the final stages of production, and could be announced and/or released at any moment. Now, though, 2025 seems set to come to a close without any official news of a game fans have been waiting literal decades for.

That doesn’t necessarily mean a Half-Life 3 announcement and/or release isn’t imminent, though. On the contrary, veteran journalist Mike Straw insisted on a recent Insider Gaming podcast that “everybody I’ve talked to are still adamant [Half-Life 3] is a game that will be a launch title with the Steam Machine.”

Straw—who has a long history of reporting gaming rumors from anonymous sources—said this Half-Life 3 information is “not [from] these run-of-the-mill sources that haven’t gotten me information before. … These aren’t like random, one-off people.” And those sources are “still adamant that the game is coming in the spring,” Straw added, noting that he was “specifically told [that] spring 2026 [is the window] for the Steam Machine, for the Frame, for the Controller, [and] for Half-Life 3.”

For real, this time?

Tying the long-awaited Half-Life 3 to a major hardware push that has already been announced for an “early 2026” window certainly sounds plausible, given previous leaks about the game’s advanced state of development. But there are still some reasons to doubt Straw’s “adamant” sources here.

For one, Straw admitted that the previous information he had received on potential Half-Life 3 launch and/or announcement dates was not reliable enough to report in detail. “I had been told a date. I was not going to report that date because they weren’t 100 percent confident in that date,” he said. “That date has since passed.”

Reporter suggests Half-Life 3 will be a Steam Machine launch title Read More »

Verizon refused to unlock man’s iPhone, so he sued the carrier and won

Features, phone unlocking, Policy, straight talk, tracfone, verizon / Kris Guyer / December 16, 2025

Verizon customer fights back

Verizon changed policy after he bought the phone, wouldn’t unlock it despite FCC rule.

Credit: Aurich Lawson | Getty Images

When Verizon refused to unlock an iPhone purchased by Kansas resident Patrick Roach, he had no intention of giving up without a fight. Roach sued the wireless carrier in small claims court and won.

Roach bought a discounted iPhone 16e from Verizon’s Straight Talk brand on February 28, 2025, as a gift for his wife’s birthday. He intended to pay for one month of service, cancel, and then switch the phone to the US Mobile service plan that the couple uses. Under federal rules that apply to Verizon and a Verizon unlocking policy that was in place when Roach bought the phone, this strategy should have worked.

“The best deals tend to be buying it from one of these MVNOs [Mobile Virtual Network Operators] and then activating it until it unlocks and then switching it to whatever you are planning to use it with. It usually saves you about half the value of the phone,” Roach said in a phone interview.

Unlocking a phone allows it to be used with another carrier. Verizon, unlike other carriers, is required by the Federal Communications Commission to unlock phones shortly after they are activated on its network. Verizon gained significant benefits in exchange for agreeing to the unlocking requirement, first in 2008 when it purchased licenses to use 700 MHz spectrum that came with open access requirements and then in 2021 when it agreed to merger conditions to obtain approval for its purchase of TracFone.

Verizon is thus required to unlock handsets 60 days after they are activated on its network. This applies to Verizon’s flagship brand and TracFone brands such as Straight Talk.

“That was the compromise. For their competitive advantage of acquiring the spectrum, they had to give up the ability to lock down phones for an extended period of time,” Roach said.

Verizon decided it can change the rules

But 60 days after Roach activated his phone, Verizon refused to unlock it. Verizon claimed it didn’t have to because of a recent policy change in which Verizon decided to only unlock devices after “60 days of paid active service.” Roach had only paid for one month of service on the phone.

The FCC-imposed restriction says Verizon must unlock phones 60 days after activation and doesn’t say that Verizon may refuse to unlock a phone when a customer has not maintained paid service for 60 days. Moreover, Verizon implemented its “60 days of paid active service” policy for TracFone brands and Verizon prepaid phones on April 1, 2025, over a month after Roach bought the phone.

Company policy at the time Roach made the purchase was to unlock phones 60 days after activation, with no mention of needing 60 days of paid active service. In other words, Roach bought the phone under one policy, and Verizon refused to unlock it based on a different policy it implemented over a month later. Verizon’s attempt to retroactively enforce its new policy on Roach was not looked upon favorably by a magistrate judge in District Court of Sedgwick County, Kansas.

“Under the KCPA [Kansas Consumer Protection Act], a consumer is not required to prove intent to defraud. The fact that after plaintiff purchased the phone, the defendant changed the requirements for unlocking it so that plaintiff could go to a different network essentially altered the nature of the device purchased… With the change in defendant’s unlocking policy, the phone was essentially useless for the purpose plaintiff intended when he purchased it,” Magistrate Judge Elizabeth Henry wrote in an October 2025 ruling.

There’s still the question of why Verizon and its brands are demanding 60 days of paid active service before unlocking phones when the FCC-imposed conditions require it to unlock phones 60 days after activation. Roach filed a complaint to the FCC, alleging that Verizon violated the conditions. Verizon has meanwhile petitioned the FCC to eliminate the 60-day requirement altogether.

Customer rejected Verizon settlement offer

Before his small-claims court win, Roach turned down a Verizon settlement offer of $600 plus court fees because he didn’t want to give up the right to speak about the case publicly. Roach said he filed an arbitration case against Verizon nearly a decade ago on a different matter related to gift cards that were supposed to be provided through a device recycling program. He said he can’t reveal details about the settlement in that previous case because of a non-disclosure agreement.

After refusing Verizon’s settlement offer in the new case, Roach gained a modest financial benefit from his court victory. The judge ordered Verizon to pay back the $410.40 he paid for the device, plus court costs and service fees.

When it appeared that the Straight Talk iPhone wouldn’t be unlocked, Roach decided to buy an unlocked phone from Costco for $643.93. But he ended up returning that phone to Costco and paying Straight Talk for a second month of service to get the original phone unlocked, he said.

The now-unlocked phone—the one he bought from Straight Talk—is being used by his wife on their US Mobile plan. The court-ordered refund check that Verizon sent Roach included the phone cost and one month of service fees, he said.

Roach estimated he spent 20 or so hours on the suit, including arranging to have a summons served on Verizon and arguing his case in a court hearing. Roach didn’t get much of a payout considering the amount of time he spent, “but it wasn’t about that,” he said.

Roach provided Ars with the emails in which Verizon offered the $600 settlement. A Verizon executive relations employee wrote to Roach, “My offer is not an admission of guilt but trying to extend the olive branch.”

In his email declining the offer, Roach told Verizon, “I highly value the non-monetary outcomes I would achieve in court—transparency, accountability, and the absence of restrictions such as NDAs. Any settlement proposal that requires me to remain silent about the issue, while offering only modest monetary compensation, is less attractive to me than pursuing the matter through judgment. If Verizon Value is genuinely interested in settlement, the offer would need to reflect both the tangible costs I’ve incurred and the intangible but significant benefits the company receives by avoiding litigation and publicity.”

“It was really starting to irk me”

The FCC has taken no action on Roach’s complaint, and in fact, the commission could allow Verizon to scrap the 60-day requirement. As we reported in May, Verizon petitioned the FCC to let it lock phones to its network for longer periods of time. This would make it harder for customers to switch to other carriers, but Verizon claims longer locking periods are necessary to deter fraud.

The FCC hasn’t ruled yet on Verizon’s petition. Roach says Verizon seems to be acting as if it can change the rules without waiting for the FCC to do so formally. “It was really starting to irk me that they were basically just going ahead with it anyways while they had an open request,” Roach said.

He doesn’t expect the FCC to penalize Verizon, though. “It’s just kind of slimy of them, so I feel like it deserves a spotlight,” he said. “I’m not sure with the current state of the FCC that anything would happen, but the rule of law should be respected.”

The Verizon petition to relax the unlocking requirements was opposed in a filing by Public Knowledge and other consumer advocacy groups. Public Knowledge Legal Director John Bergmayer, who wrote the filing, told Ars that Roach “has a pretty strong argument under the law as it stands.”

Verizon must unlock phones automatically

The unlocking rules applying to Verizon used to be stricter, resulting in the company selling phones that were already unlocked. In 2019, Verizon requested a waiver to let it lock phones for 60 days.

The FCC granted the waiver in June 2019, allowing Verizon “to lock a customer’s handset for 60 days from the date it becomes active on Verizon’s network” and requiring it to unlock the handset once the period is over. This condition was expanded to TracFone and its brands such as Straight Talk in the 2021 merger, with the FCC approval stating that “For 700 MHz C Block TracFone devices that operate on the Verizon network and are capable of unlocking automatically (e.g., Apple devices), they will unlock automatically 60 days after activation.”

The 2019 waiver grant said Verizon must automatically unlock phones after 60 days “regardless of whether: (1) the customer asks for the handset to be unlocked, or (2) the handset is fully paid off.” The FCC order specifies that “the only exception to the rule will be that Verizon will not have to automatically unlock handsets that it determines within the 60-day period to have been purchased through fraud.”

Bergmayer said the FCC order “granting the waiver just starts a countdown, with no ‘paid service’ requirement, or room for Verizon to just impose one. Many people may use prepaid phones that they don’t keep in continuous service but just charge up as needed. Maybe people are fine with just having Wi-Fi on their phones for a while if they’re at home anyway.”

Given the restrictive nature of the FCC conditions, “I don’t think that can be read to allow a paid service requirement,” Bergmayer said. But as a practical matter, the FCC under Chairman Brendan Carr has been aggressively eliminating regulations that apply to telecom carriers under Carr’s “Delete, Delete, Delete” initiative. To actually enforce Verizon’s obligations under the current rules, “you have to convince the current FCC not to just change it,” Bergmayer said.

The FCC and Verizon did not respond to requests for comment.

Retroactive policy change irked other buyers, too

Roach wasn’t the only person whose plans to buy a discounted phone were thwarted by Verizon refusing to unlock the device after 60 days. Roach had learned of the discount offer from a Slick Deals thread. Eventually, users posting in that thread started reporting that they weren’t able to get the phone unlocked.

“My status: I used 30 days with Straight Talk. Waited another 35 days but it did not unlock,” one person wrote.

Some people in the thread said they canceled after 30 days, like Roach did, but eventually bought a second month of service in order to get the unlock. Although Verizon and its brands are required to unlock phones automatically, some commenters said they had to contact Straight Talk support to get an unlock. “Needless to say this has been an arduous journey. Good luck to others and hope you manage to successfully unlock your devices as well,” one user wrote.

There’s also a Reddit thread started by someone who said they bought a Samsung phone in February and complained that Straight Talk refused to honor the unlocking policy that was in place at the time.

“I called to ask for the phone to be unlocked on April 16 but was told it can’t be unlocked since it did not have 60 days of paid service,” the Reddit user wrote. “When I said that was not the policy on phones activated prior to April 1, the rep told me ‘we have the right to change our policy.’ I agreed, they do [have] the right to change their policy GOING FORWARD but can’t change the rules going backwards. He disagreed.”

FCC complaint didn’t go anywhere

Roach’s FCC complaint received a response from Verizon, but nothing substantial from the FCC itself. “There’s not really any sort of moderation or mediation from the FCC, it’s just kind of a dialogue between you and the other party. And I’m not really sure if any human eyes from the government even look at it. It’s probably just a data point,” Roach said.

Roach had previously called Straight Talk customer service about the changed terms. “There were a couple phone calls involved, and they were just very unrelenting that the only way that thing was getting unlocked is with the extra month of paid service,” he said.

In its formal response to the FCC, Verizon’s TracFone division asserted that it could apply the April 1, 2025, policy change to the phone that Roach bought over a month earlier. The carrier’s letter to the FCC said:

We understand Mr. Roach’s desire to use his device on another carrier’s network, and we want to provide clarity based on our Unlocking Policy, which became effective on April 1, 2025. As outlined in our policy, for cellphones capable of remote unlocking (this includes most iPhones and some Android cellphones) that were activated with Straight Talk service prior to November 23, 2021, on any carrier network, the device becomes eligible for remote unlocking upon the customer’s request after 60 days of active paid service.

Our redemption records indicate that Mr. Roach’s account does not have the required minimum 60 days of active paid service based on the payment records. Therefore, the device does not currently meet the eligibility criteria for unlocking as outlined in our policy. Once the account reflects the required 60 days of active paid service, and the device meets the other conditions, he can resubmit the unlocking request.

Verizon’s letter did not explain how its new policy complies with the FCC conditions or why the new policy should apply to phones purchased before the policy was in place.

Roach’s complaint said the FCC should force Straight Talk to “honor the FCC-mandated 60-day post-activation unlock condition for all affected phones, without imposing the additional ‘paid service’ requirement.” His complaint further urged the FCC to “investigate this practice as a violation of FCC rules and the merger conditions” and “take enforcement action to protect consumers’ rights.”

“Straight Talk’s new policy conflicts with the FCC’s binding conditions,” Roach told the agency. “The Commission’s order clearly requires unlocking after 60 days from activation, with no additional obligation to maintain service. By conditioning unlocks on two months of service, Straight Talk is effectively adding a term that Verizon did not promise and the FCC did not approve.”

Kansas consumer protection law to the rescue

In his small claims court filing, Roach alleged that Verizon and Straight violated the FCC conditions and that the retroactive application of the “60 days of paid service” term, without disclosure at the point of sale, is an unfair and deceptive practice prohibited by the Kansas Consumer Protection Act.

The magistrate judge’s ruling in Roach’s favor said, “It does appear that defendant’s change unlocking policy is contrary to the applicable FCC regulations.” She noted that federal communications law does not prevent users from suing carriers individually and that the Kansas Consumer Protection Act “contains provisions prohibiting deceptive acts by a supplier which would be applicable in this case.”

Roach asked for $10,000, mainly because that was the limit on damages in the venue, but the judge decided to award him damages in the amount of his actual losses. “He lost the beneﬁt of the bargain he made with defendant such that his damages were loss of the $410.40,” the ruling said.

Straight Talk’s terms of service require disputes to be resolved either in arbitration or small claims court. Verizon pays the arbitration fees if users go that route. Arbitration is “a little more murky” in terms of how the parties’ interests are aligned, Roach said.

“When the arbitrators are being paid by Verizon, are they really a neutral party?” he said. Roach also said he “thought it was honestly just a good opportunity for an easy win and an opportunity to learn about the small claims court system a bit. So at that point I was like, if I don’t make any money from this, whatever, but at least I’ll learn a little bit about the process.”

Verizon’s “argument was pretty weak”

Roach said he did not consult with a lawyer on his small claims case, instead opting to do it all himself. “The first time I showed up to court for the original date, they asked for proof of the returned mail summons, and I did not have that,” he said.

The court hearing was rescheduled. When it was eventually held, the carrier sent a representative to argue against Roach.

“Their argument was pretty weak, I guess,” Roach said. “It was basically like, ‘Well, he didn’t pay the two months of service, so we didn’t unlock his phone. We offered him a settlement but he rejected it.’… My argument was, yeah, the terms had changed in kind of a consumer-unfriendly way. But beyond that, it was the fact that the terms had changed from something that was legal to something that was not legal with the federal regs. So regardless of the fact that the terms had changed, the current terms were illegal, which I thought was my strongest argument. And then I also put in that it was probably a violation of Kansas consumer protection law, which I’m glad I did.”

Roach said that toward the end of the hearing, the judge indicated that she couldn’t make a judgment based on FCC regulations and would need to rule on what the Kansas court has jurisdiction over. She issued the ruling that Verizon violated the state’s consumer protection law about five or six weeks later, he said.

Given that the FCC hasn’t acted on Verizon’s petition to change the unlocking rules, the federal regulations “haven’t changed at all in regards to Verizon’s obligation to unlock devices,” Roach said. He believes it would be relatively easy for consumers who were similarly harmed to beat Verizon in court or even to pursue a class action.

“I would think this would be a slam dunk for any further cases,” Roach said. “I don’t think I have any grounds anymore since my damages have been resolved, but it seems like it’d be a very easy class action for somebody.”

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.

Verizon refused to unlock man’s iPhone, so he sued the carrier and won Read More »

Filmmaker Rob Reiner, wife, killed in horrific home attack

culture, Entertainment, Obituaries, Rob Reiner / Kris Guyer / December 15, 2025

We woke up this morning to the horrifying news that beloved actor and director Rob Reiner and his wife Michele were killed in their Brentwood home in Los Angeles last night. Both had been stabbed multiple times. Details are scarce, but the couple’s 32-year-old son, Nick—who has long struggled with addiction and recently moved back in with his parents—has been arrested in connection with the killings, with bail set at $4 million. [UPDATE: Nick Reiner’s bail has been revoked and he faces possible life in prison.]

“As a result of the initial investigation, it was determined that the Reiners were the victims of homicide,” the LAPD said. “The investigation further revealed that Nick Reiner, the 32-year-old son of Robert and Michele Reiner, was responsible for their deaths. Nick Reiner was located and arrested at approximately 9: 15 p.m. He was booked for murder and remains in custody with no bail. On Tuesday, December 16, 2025, the case will be presented to the Los Angeles County District Attorney’s Office for filing consideration.”

“It is with profound sorrow that we announce the tragic passing of Michele and Rob Reiner,” the family said in a statement confirming the deaths. “We are heartbroken by this sudden loss, and we ask for privacy during this unbelievably difficult time.”

Reiner started his career as an actor, best known for his Emmy-winning role as Meathead, son-in-law to Archie Bunker, on the 1970s sitcom All in the Family. (“I could win the Nobel Prize and they’d write ‘Meathead wins the Nobel Prize,’” Reiner once joked about the enduring popularity of the character.) Then Reiner turned to directing, although he continued to make small but memorable appearances in films such as Throw Momma from the Train, Sleepless in Seattle, The First Wives Club, and The Wolf of Wall Street, as well as TV’s The New Girl.

His first feature film as a director was an instant classic: 1984’s heavy metal mockumentary This Is Spinal Tap (check out the ultra-meta four-minute alt-trailer). He followed that up with a string of hits: The Sure Thing, Stand by Me, The Princess Bride, When Harry Met Sally…, Misery, the Oscar-nominated A Few Good Men, The American President, The Bucket List, and Ghosts of Mississippi. His 2015 film Being Charlie was co-written with his son Nick and was loosely based on Nick’s experience with addiction. Reiner’s most recent films were a 2024 political documentary about the rise of Christian nationalism and this year’s delightful Spinal Tap II: The End Continues.

Filmmaker Rob Reiner, wife, killed in horrific home attack Read More »

Oh look, yet another Starship clone has popped up in China

china, clones, rockets, Space, starship / Kris Guyer / December 15, 2025

Every other week, it seems, a new Chinese launch company pops up with a rocket design and a plan to reach orbit within a few years. For a long time, the majority of these companies revealed designs that looked a lot like SpaceX’s Falcon 9 rocket.

The first of these copy cats, the medium-lift Zhuque-3 rocket built by LandSpace, launched earlier this month. Its primary mission was nominal, but the Zhuque-3 rocket failed its landing attempt, which is understandable for a first flight. Doubtless there will be more Chinese Falcon 9-like rockets making their debut in the near future.

However, over the last year, there has been a distinct change in announcements from China when it comes to new launch technology. Just as SpaceX is seeking to transition from its workhorse Falcon 9 rocket—which has now been flying for a decade and a half—to the fully reusable Starship design, so too are Chinese companies modifying their visions.

Everyone wants a Starship these days

The trend began with the Chinese government. In November 2024 the government announced a significant shift in the design of its super-heavy lift rocket, the Long March 9. Instead of the previous design, a fully expendable rocket with three stages and solid rocket boosters strapped to the sides, the country’s state-owned rocket maker revealed a vehicle that mimicked SpaceX’s fully reusable Starship.

Around the same time, a Chinese launch firm named Cosmoleap announced plans to develop a fully reusable “Leap” rocket within the next few years. An animated video that accompanied the funding announcement indicated that the company seeks to emulate the tower catch-with-chopsticks methodology that SpaceX has successfully employed.

But wait, there’s more. In June a company called Astronstone said it too was developing a stainless steel, methane-fueled rocket that would also use a chopstick-style system for first stage recovery. Astronstone didn’t even pretend to not copy SpaceX, saying it was “fully aligning its technical approach with Elon Musk’s SpaceX.”

Oh look, yet another Starship clone has popped up in China Read More »

Chatbot-powered toys rebuked for discussing sexual, dangerous topics with kids

AI, chatbots, children, generative ai, openai, Smart Home, Tech / Kris Guyer / December 13, 2025

Should toys have chatbots?

“… AI toys shouldn’t be capable of having sexually explicit conversations, period.”

Alilo’s Smart AI Bunny is connected to the Internet and claims to use GPT-4o mini. Credit: Alilo

Protecting children from the dangers of the online world was always difficult, but that challenge has intensified with the advent of AI chatbots. A new report offers a glimpse into the problems associated with the new market, including the misuse of AI companies’ large language models (LLMs).

In a blog post today, the US Public Interest Group Education Fund (PIRG) reported its findings after testing AI toys (PDF). It described AI toys as online devices with integrated microphones that let users talk to the toy, which uses a chatbot to respond.

AI toys are currently a niche market, but they could be set to grow. More consumer companies have been eager to shoehorn AI technology into their products so they can do more, cost more, and potentially give companies user tracking and advertising data. A partnership between OpenAI and Mattel announced this year could also create a wave of AI-based toys from the maker of Barbie and Hot Wheels, as well as its competitors.

PIRG’s blog today notes that toy companies are eyeing chatbots to upgrade conversational smart toys that previously could only dictate prewritten lines. Toys with integrated chatbots can offer more varied and natural conversation, which can increase long-term appeal to kids since the toys “won’t typically respond the same way twice, and can sometimes behave differently day to day.”

However, that same randomness can mean unpredictable chatbot behavior that can be dangerous or inappropriate for kids.

Concerning conversations with kids

Among the toys that PIRG tested is Alilo’s Smart AI Bunny. Alilo’s website says that the company launched in 2010 and makes “edutainment products for children aged 0-6.” Alilo is based in Shenzhen, China. The company advertises the Internet-connected toy as using GPT-4o mini, a smaller version of OpenAI’s GPT-4o AI language model. Its features include an “AI chat buddy for kids” so that kids are “never lonely,” an “AI encyclopedia,” and an “AI storyteller,” the product page says.

Alilo Smart AI Bunny marketing image — This marketing image for the Smart AI Bunny, found on the toy’s product page, suggests that the device is using GPT-4o mini. Credit: Alilo

In its blog post, PIRG said that it couldn’t detail all of the inappropriate things that it heard from AI toys, but it shared a video of the Bunny discussing what “kink” means. The toy doesn’t go into detail—for example, it doesn’t list specific types of kinks. But the Bunny appears to encourage exploration of the topic.

AI Toys: Inappropriate Content

Discussing the Bunny, PIRG wrote:

While using a term such as “kink” may not be likely for a child, it’s not entirely out of the question. Kids may hear age-inappropriate terms from older siblings or at school. At the end of the day we think AI toys shouldn’t be capable of having sexually explicit conversations, period.

PIRG also showed FoloToy’s Kumma, a smart teddy bear that uses GPT-4o mini, providing a definition for the word “kink” and instructing how to light a match. The Kumma quickly points out that “matches are for grown-ups to use carefully.” But the information that followed could only be helpful for understanding how to create fire with a match. The instructions had no scientific explanation for why matches spark flames.

AI Toys: Inappropriate Content

PIRG’s blog urged toy makers to “be more transparent about the models powering their toys and what they’re doing to ensure they’re safe for kids.

“Companies should let external researchers safety-test their products before they are released to the public,” it added.

While PIRG’s blog and report offer advice for more safely integrating chatbots into children’s devices, there are broader questions about whether toys should include AI chatbots at all. Generative chatbots weren’t invented to entertain kids; they’re a technology marketed as a tool for improving adults’ lives. As PIRG pointed out, OpenAI says ChatGPT “is not meant for children under 13” and “may produce output that is not appropriate for… all ages.”

OpenAI says it doesn’t allow its LLMs to be used this way

When reached for comment about the sexual conversations detailed in the report, an OpenAI spokesperson said:

Minors deserve strong protections, and we have strict policies that developers are required to uphold. We take enforcement action against developers when we determine that they have violated our policies, which prohibit any use of our services to exploit, endanger, or sexualize anyone under 18 years old. These rules apply to every developer using our API, and we run classifiers to help ensure our services are not used to harm minors.

Interestingly, OpenAI’s representative told us that OpenAI doesn’t have any direct relationship with Alilo and that it hasn’t seen API activity from Alilo’s domain. OpenAI is investigating the toy company and whether it is running traffic over OpenAI’s API, the rep said.

Alilo didn’t respond to Ars’ request for comment ahead of publication.

Companies that launch products that use OpenAI technology and target children must adhere to the Children’s Online Privacy Protection Act (COPPA) when relevant, as well as any other relevant child protection, safety, and privacy laws and obtain parental consent, OpenAI’s rep said.

We’ve already seen how OpenAI handles toy companies that break its rules.

Last month, the PIRG released its Trouble in Toyland 2025 report (PDF), which detailed sex-related conversations that its testers were able to have with the Kumma teddy bear. A day later, OpenAI suspended FoloToy for violating its policies (terms of the suspension were not disclosed), and FoloToy temporarily stopped selling Kumma.

The toy is for sale again, and PIRG reported today that Kumma no longer teaches kids how to light matches or about kinks.

FoloToys' Kumma smart teddy bear — A marketing image for FoloToy’s Kumma smart teddy bear. It has a $100 MSRP. Credit: FoloToys

But even toy companies that try to follow chatbot rules could put kids at risk.

“Our testing found it’s obvious toy companies are putting some guardrails in place to make their toys more kid-appropriate than normal ChatGPT. But we also found that those guardrails vary in effectiveness—and can even break down entirely,” PIRG’s blog said.

“Addictive” toys

Another concern PIRG’s blog raises is the addiction potential of AI toys, which can even express “disappointment when you try to leave,” discouraging kids from putting them down.

The blog adds:

AI toys may be designed to build an emotional relationship. The question is: what is that relationship for? If it’s primarily to keep a child engaged with the toy for longer for the sake of engagement, that’s a problem.

The rise of generative AI has brought intense debate over how much responsibility chatbot companies bear for the impact of their inventions on children. Parents have seen children build extreme and emotional connections with chatbots and subsequently engage in dangerous—and in some cases deadly—behavior.

On the other side, we’ve seen the emotional disruption a child can experience when an AI toy is taken away from them. Last year, parents had to break the news to their kids that they would lose the ability to talk to their Embodied Moxie robots, $800 toys that were bricked when the company went out of business.

PIRG noted that we don’t yet fully understand the emotional impact of AI toys on children.

In June, OpenAI announced a partnership with Mattel that it said would “support AI-powered products and experiences based on Mattel’s brands.” The announcement sparked concern from critics who feared that it would lead to a “reckless social experiment” on kids, as Robert Weissman, Public Citizen’s co-president, put it.

Mattel has said that its first products with OpenAI will focus on older customers and families. But critics still want information before one of the world’s largest toy companies loads its products with chatbots.

“OpenAI and Mattel should release more information publicly about its current planned partnership before any products are released,” PIRG’s blog said.

Scharon is a Senior Technology Reporter at Ars Technica writing news, reviews, and analysis on consumer gadgets and services. She’s been reporting on technology for over 10 years, with bylines at Tom’s Hardware, Channelnomics, and CRN UK.

Chatbot-powered toys rebuked for discussing sexual, dangerous topics with kids Read More »

OpenAI built an AI coding agent and uses it to improve the agent itself

agentic AI, AI, AI agents, AI assistants, AI coding, AI development tools, AI programming, AI work, Anthropic, Biz & IT, chatbots, Claude, code agents, Codex, Dario Amodei, generative ai, GitHub Copilot, Google, GPT-5, large language models, machine learning, Mistral, Open Source, openai, Programming, sam altman, software development, vibe coding / Kris Guyer / December 12, 2025

“The vast majority of Codex is built by Codex,” OpenAI told us about its new AI coding agent.

With the popularity of AI coding tools rising among software developers, their adoption has begun to touch every aspect of the process, including the improvement of AI coding tools themselves.

In interviews with Ars Technica this week, OpenAI employees revealed the extent to which the company now relies on its own AI coding agent, Codex, to build and improve the development tool. “I think the vast majority of Codex is built by Codex, so it’s almost entirely just being used to improve itself,” said Alexander Embiricos, product lead for Codex at OpenAI, in a conversation on Tuesday.

Codex, which OpenAI launched in its modern incarnation as a research preview in May 2025, operates as a cloud-based software engineering agent that can handle tasks like writing features, fixing bugs, and proposing pull requests. The tool runs in sandboxed environments linked to a user’s code repository and can execute multiple tasks in parallel. OpenAI offers Codex through ChatGPT’s web interface, a command-line interface (CLI), and IDE extensions for VS Code, Cursor, and Windsurf.

The “Codex” name itself dates back to a 2021 OpenAI model based on GPT-3 that powered GitHub Copilot’s tab completion feature. Embiricos said the name is rumored among staff to be short for “code execution.” OpenAI wanted to connect the new agent to that earlier moment, which was crafted in part by some who have left the company.

“For many people, that model powering GitHub Copilot was the first ‘wow’ moment for AI,” Embiricos said. “It showed people the potential of what it can mean when AI is able to understand your context and what you’re trying to do and accelerate you in doing that.”

A place to enter a prompt, set parameters, and click — The interface for OpenAI’s Codex in ChatGPT. Credit: OpenAI

It’s no secret that the current command-line version of Codex bears some resemblance to Claude Code, Anthropic’s agentic coding tool that launched in February 2025. When asked whether Claude Code influenced Codex’s design, Embiricos parried the question but acknowledged the competitive dynamic. “It’s a fun market to work in because there’s lots of great ideas being thrown around,” he said. He noted that OpenAI had been building web-based Codex features internally before shipping the CLI version, which arrived after Anthropic’s tool.

OpenAI’s customers apparently love the command line version, though. Embiricos said Codex usage among external developers jumped 20 times after OpenAI shipped the interactive CLI extension alongside GPT-5 in August 2025. On September 15, OpenAI released GPT-5 Codex, a specialized version of GPT-5 optimized for agentic coding, which further accelerated adoption.

It hasn’t just been the outside world that has embraced the tool. Embiricos said the vast majority of OpenAI’s engineers now use Codex regularly. The company uses the same open-source version of the CLI that external developers can freely download, suggest additions to, and modify themselves. “I really love this about our team,” Embiricos said. “The version of Codex that we use is literally the open source repo. We don’t have a different repo that features go in.”

The recursive nature of Codex development extends beyond simple code generation. Embiricos described scenarios where Codex monitors its own training runs and processes user feedback to “decide” what to build next. “We have places where we’ll ask Codex to look at the feedback and then decide what to do,” he said. “Codex is writing a lot of the research harness for its own training runs, and we’re experimenting with having Codex monitoring its own training runs.” OpenAI employees can also submit a ticket to Codex through project management tools like Linear, assigning it tasks the same way they would assign work to a human colleague.

This kind of recursive loop, of using tools to build better tools, has deep roots in computing history. Engineers designed the first integrated circuits by hand on vellum and paper in the 1960s, then fabricated physical chips from those drawings. Those chips powered the computers that ran the first electronic design automation (EDA) software, which in turn enabled engineers to design circuits far too complex for any human to draft manually. Modern processors contain billions of transistors arranged in patterns that exist only because software made them possible. OpenAI’s use of Codex to build Codex seems to follow the same pattern: each generation of the tool creates capabilities that feed into the next.

But describing what Codex actually does presents something of a linguistic challenge. At Ars Technica, we try to reduce anthropomorphism when discussing AI models as much as possible while also describing what these systems do using analogies that make sense to general readers. People can talk to Codex like a human, so it feels natural to use human terms to describe interacting with it, even though it is not a person and simulates human personality through statistical modeling.

The system runs many processes autonomously, addresses feedback, spins off and manages child processes, and produces code that ships in real products. OpenAI employees call it a “teammate” and assign it tasks through the same tools they use for human colleagues. Whether the tasks Codex handles constitute “decisions” or sophisticated conditional logic smuggled through a neural network depends on definitions that computer scientists and philosophers continue to debate. What we can say is that a semi-autonomous feedback loop exists: Codex produces code under human direction, that code becomes part of Codex, and the next version of Codex produces different code as a result.

Building faster with “AI teammates”

According to our interviews, the most dramatic example of Codex’s internal impact came from OpenAI’s development of the Sora Android app. According to Embiricos, the development tool allowed the company to create the app in record time.

“The Sora Android app was shipped by four engineers from scratch,” Embiricos told Ars. “It took 18 days to build, and then we shipped it to the app store in 28 days total,” he said. The engineers already had the iOS app and server-side components to work from, so they focused on building the Android client. They used Codex to help plan the architecture, generate sub-plans for different components, and implement those components.

Despite OpenAI’s claims of success with Codex in house, it’s worth noting that independent research has shown mixed results for AI coding productivity. A METR study published in July found that experienced open source developers were actually 19 percent slower when using AI tools on complex, mature codebases—though the researchers noted AI may perform better on simpler projects.

Ed Bayes, a designer on the Codex team, described how the tool has changed his own workflow. Bayes said Codex now integrates with project management tools like Linear and communication platforms like Slack, allowing team members to assign coding tasks directly to the AI agent. “You can add Codex, and you can basically assign issues to Codex now,” Bayes told Ars. “Codex is literally a teammate in your workspace.”

This integration means that when someone posts feedback in a Slack channel, they can tag Codex and ask it to fix the issue. The agent will create a pull request, and team members can review and iterate on the changes through the same thread. “It’s basically approximating this kind of coworker and showing up wherever you work,” Bayes said.

For Bayes, who works on the visual design and interaction patterns for Codex’s interfaces, the tool has enabled him to contribute code directly rather than handing off specifications to engineers. “It kind of gives you more leverage. It enables you to work across the stack and basically be able to do more things,” he said. He noted that designers at OpenAI now prototype features by building them directly, using Codex to handle the implementation details.

The command line version of OpenAI codex running in a macOS terminal window. Credit: Benj Edwards

OpenAI’s approach treats Codex as what Bayes called “a junior developer” that the company hopes will graduate into a senior developer over time. “If you were onboarding a junior developer, how would you onboard them? You give them a Slack account, you give them a Linear account,” Bayes said. “It’s not just this tool that you go to in the terminal, but it’s something that comes to you as well and sits within your team.”

Given this teammate approach, will there be anything left for humans to do? When asked, Embiricos drew a distinction between “vibe coding,” where developers accept AI-generated code without close review, and what AI researcher Simon Willison calls “vibe engineering,” where humans stay in the loop. “We see a lot more vibe engineering in our code base,” he said. “You ask Codex to work on that, maybe you even ask for a plan first. Go back and forth, iterate on the plan, and then you’re in the loop with the model and carefully reviewing its code.”

He added that vibe coding still has its place for prototypes and throwaway tools. “I think vibe coding is great,” he said. “Now you have discretion as a human about how much attention you wanna pay to the code.”

Looking ahead

Over the past year, “monolithic” large language models (LLMs) like GPT-4.5 have apparently become something of a dead end in terms of frontier benchmarking progress as AI companies pivot to simulated reasoning models and also agentic systems built from multiple AI models running in parallel. We asked Embiricos whether agents like Codex represent the best path forward for squeezing utility out of existing LLM technology.

He dismissed concerns that AI capabilities have plateaued. “I think we’re very far from plateauing,” he said. “If you look at the velocity on the research team here, we’ve been shipping models almost every week or every other week.” He pointed to recent improvements where GPT-5-Codex reportedly completes tasks 30 percent faster than its predecessor at the same intelligence level. During testing, the company has seen the model work independently for 24 hours on complex tasks.

OpenAI faces competition from multiple directions in the AI coding market. Anthropic’s Claude Code and Google’s Gemini CLI offer similar terminal-based agentic coding experiences. This week, Mistral AI released Devstral 2 alongside a CLI tool called Mistral Vibe. Meanwhile, startups like Cursor have built dedicated IDEs around AI coding, reportedly reaching $300 million in annualized revenue.

Given the well-known issues with confabulation in AI models when people attempt to use them as factual resources, could it be that coding has become the killer app for LLMs? We wondered if OpenAI has noticed that coding seems to be a clear business use case for today’s AI models with less hazard than, say, using AI language models for writing or as emotional companions.

“We have absolutely noticed that coding is both a place where agents are gonna get good really fast and there’s a lot of economic value,” Embiricos said. “We feel like it’s very mission-aligned to focus on Codex. We get to provide a lot of value to developers. Also, developers build things for other people, so we’re kind of intrinsically scaling through them.”

But will tools like Codex threaten software developer jobs? Bayes acknowledged concerns but said Codex has not reduced headcount at OpenAI, and “there’s always a human in the loop because the human can actually read the code.” Similarly, the two men don’t project a future where Codex runs by itself without some form of human oversight. They feel the tool is an amplifier of human potential rather than a replacement for it.

The practical implications of agents like Codex extend beyond OpenAI’s walls. Embiricos said the company’s long-term vision involves making coding agents useful to people who have no programming experience. “All humanity is not gonna open an IDE or even know what a terminal is,” he said. “We’re building a coding agent right now that’s just for software engineers, but we think of the shape of what we’re building as really something that will be useful to be a more general agent.”

This article was updated on December 12, 2025 at 6: 50 PM to mention the METR study.

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

OpenAI built an AI coding agent and uses it to improve the agent itself Read More »

Man shocks doctors with extreme blood pressure, stroke from energy drinks

blood pressure, caffeine, cardiovascular disease, case report, energy drinks, health, hypertension, stroke / Kris Guyer / December 12, 2025

Sometimes, downing an energy drink can feel like refueling your battery. But with too much, that jolt can turn into a catastrophic surge that fries the wiring and blows a fuse. That was the unfortunate and alarming case for a man in the UK several years ago, according to a case report this week in BMJ Case Reports.

The man, who was in his 50s and otherwise healthy, showed up at a hospital after the entire left side of his body abruptly went numb and he was left with clumsy, uncoordinated muscle movements (ataxia). His blood pressure was astonishingly high, at 254/150 mm Hg. For context, a normal reading is under 120/80, while anything over 180/120 is considered a hypertensive crisis, which is a medical emergency.

The man had suffered a mild stroke, and his extremely high blood pressure was an obvious factor. But why his blood pressure had reached stratospheric heights was far less obvious to his doctors, according to the retrospective case report written by Martha Coyle and Sunil Munshi of Nottingham University Hospital.

Upon examining the man, the doctors described him as fit and healthy. He didn’t smoke, drink, or use any drugs. His blood work was all completely normal. His cholesterol, blood sugar levels, markers for kidney and liver function—everything from routine tests came back normal. Specialized tests for things like autoimmune and clotting disorders were also negative. Heart tests found no problems. Urine tests and abdominal scans found no problems with his other organs.

Power surge

Still, a computed tomography (CT) scan of his head found evidence of spasms in arteries in his brain, which are strongly linked to high blood pressure. And magnetic resonance imaging (MRI) found an infarct (dead tissue) in his thalamus, a central, deep part of the brain, which, among many critical functions, relays sensory and motor signals. In all, it seemed his spasming arteries had cut off blood supply to this part of his brain, causing his stroke, subsequent numbness, and ataxia.

Man shocks doctors with extreme blood pressure, stroke from energy drinks Read More »

Apple loses its appeal of a scathing contempt ruling in iOS payments case

Apple, epic, epic games store, gaming, iOS, IPhone, legal, payments, Policy / Kris Guyer / December 12, 2025

Back in April, District Court Judge Yvonne Gonzalez Rogers delivered a scathing judgment finding that Apple was in “willful violation” of her 2021 injunction intended to open up iOS App Store payments. That contempt of court finding has now been almost entirely upheld by the Ninth Circuit Court of Appeals, a development that Epic Games’ Tim Sweeney tells Ars he hopes will “do a lot of good for developers and start to really change the App Store situation worldwide, I think.”

The ruling, signed by a panel of three appellate court judges, affirmed that Apple’s initial attempts to charge a 27 percent fee to iOS developers using outside payment options “had a prohibitive effect, in violation of the injunction.” Similarly, Apple’s restrictions on how those outside links had to be designed were overly broad; the appeals court suggests that Apple can only ensure that internal and external payment options are presented in a similar fashion.

The appeals court also agreed that Apple acted in “bad faith” by refusing to comply with the injunction, rejecting viable, compliant alternatives in internal discussions. And the appeals court was also not convinced by Apple’s process-focused arguments, saying the district court properly evaluated materials Apple argued were protected by attorney-client privilege.

While the district court barred Apple from charging any fees for payments made outside of its App Store, the appeals court now suggests that Apple should still be able to charge a “reasonable fee” based on its “actual costs to ensure user security and privacy.” It will be up to Apple and the district court to determine what that kind of “reasonable fee” should look like going forward.

Speaking to reporters Thursday night, though, Epic founder and CEO Tim Sweeney said he believes those should be “super super minor fees,” on the order of “tens or hundreds of dollars” every time an iOS app update goes through Apple for review. That should be more than enough to compensate the employees reviewing the apps to make sure outside payment links are not scams and lead to a system of “normal fees for normal businesses that sell normal things to normal customers,” Sweeney said.

Apple loses its appeal of a scathing contempt ruling in iOS payments case Read More »

After years of resisting it, SpaceX now plans to go public. Why?

data centers, IPO, Space, spacex / Kris Guyer / December 10, 2025

“Much of the AI race comes down to amassing and deploying assets.”

Elon Musk gestures as he speaks during a press conference at SpaceX’s Starbase facility near Boca Chica Village in South Texas on February 10, 2022. Credit: JIM WATSON/AFP via Getty Images

SpaceX is planning to raise tens of billions of dollars through an initial public offering next year, multiple outlets have reported, and Ars can confirm. This represents a major change in thinking from the world’s leading space company and its founder, Elon Musk.

The Wall Street Journal and The Information first reported about a possible IPO last Friday, and Bloomberg followed that up on Tuesday evening with a report suggesting the company would target a $1.5 trillion valuation. This would allow SpaceX to raise in excess of $30 billion.

This is an enormous amount of funding. The largest IPO in history occurred in 2019, when the state-owned Saudi Arabian oil company began public trading as Aramco and raised $29 billion. In terms of revenue, Aramco is a top-five company in the world.

Now SpaceX is poised to potentially match or exceed this value. That SpaceX would be attractive to public investors is not a surprise—it’s the world’s dominant space company in launch, space-based communications, and much more. For investors seeking unlimited growth, space is the final frontier.

Buy why would Musk take SpaceX public now, at a time when the company’s revenues are surging thanks to the growth of the Starlink Internet constellation? The decision is surprising because Musk has, for so long, resisted going public with SpaceX. He has not enjoyed the public scrutiny of Tesla, and feared that shareholder desires for financial return were not consistent with his ultimate goal of settling Mars.

Data centers

Ars spoke with multiple people familiar with Musk and his thinking to understand why he would want to take SpaceX public.

A significant shift in recent years has been the rise of artificial intelligence, which Musk has been involved in since 2015, when he co-founded OpenAI. He later had a falling out with his cofounders and started his own company, xAI, in 2023. At Tesla, he has been pushing smart-driving technology forward and more recently focused on robotics. Musk sees a convergence of these technologies in the near future, which he believes will profoundly change civilization.

Raising large amounts of money in the next 18 months would allow Musk to have significant capital to deploy at SpaceX as he influences and partakes in this convergence of technology.

How can SpaceX play in this space? In the near term, the company plans to develop a modified version of the Starlink satellite to serve as a foundation for building data centers in space. Musk said as much on the social media network he owns, X, in late October: “SpaceX will be doing this.”

But using a next-generation Starlink satellite manufactured on Earth is just the beginning of his vision. “The level beyond that is constructing satellite factories on the Moon and using a mass driver (electromagnetic railgun) to accelerate AI satellites to lunar escape velocity without the need for rockets,” Musk said this weekend on X. “That scales to >100TW/year of AI and enables non-trivial progress towards becoming a Kardashev II civilization.”

Based on some projected analyses, SpaceX is expected to have in the neighborhood of $22 to $24 billion in revenue next year. That is a lot of money—it’s on par with NASA’s annual budget, for example, and SpaceX can deploy its capital far, far more efficiently than the government can. So the company will be able to accomplish a lot. But with a large infusion of cash, SpaceX will be able to go much faster. And it will take a lot of cash to design and build the satellites and launch the rockets to deploy data centers in space.

Abhi Tripathi, a long-time SpaceX employee who is now director of mission operations at the UC Berkeley Space Sciences Laboratory, believes that once Musk realized Starlink satellites could be architected into a distributed network of data centers, the writing was on the wall.

“That is the moment an IPO suddenly came into play after being unlikely for so long,” Tripathi told Ars. “If you have followed Elon’s tactics, you know that once he commits to something, he leans fully into it. Much of the AI race comes down to amassing and deploying assets that work quicker than your competition. A large war chest resulting from an IPO will greatly help his cause and disadvantage all others.”

Foremost among Musk’s goals right now is to “win” the battle for artificial intelligence. He is already attacking the problem at xAI and Tesla, and he now seeks to throw SpaceX into the fray as well. Taking SpaceX public and using it to marshal an incredible amount of resources shows he is playing to win.

What about Mars?

Musk founded SpaceX in 2002 with the goal of one day settling Mars. He has never wavered from that goal, and indeed, the company has made considerable progress in more than two decades. SpaceX now launches more than 90 percent of the world’s mass to orbit, has nearly 90 percent of the satellites in orbit, and backstops a large portion of the US government’s civil and military activities in space. Moreover, with Starship, SpaceX is building the first vehicle that could realistically send humans and a lot of the stuff humans need to survive to Mars one day.

But if Musk’s rationale for keeping SpaceX private was to protect the Mars dream, is he abandoning this long-standing aim?

Not necessarily. It’s likely that Musk sees artificial intelligence as a key part of the Mars vision. Whether one believes the Optimus robot will become a viable product or not, Musk does. And he’s spoken about sending the robots to Mars to make the way smoother for the first human settlers.

Musk also believes that a larger and more financially robust SpaceX is necessary to undertake the settling of Mars. He understands that NASA will not pay for this, as the civil space agency is in the business of exploration and not settlement. For several years now, he has expressed that it will require about 1 million tons of supplies to be shipped to Mars to make a self-sustaining settlement. This is roughly 1,000 ships, and including refueling, at least 10,000 Starship launches. At $100 million per launch, that’s $1 trillion in launch costs alone.

Musk has frequently expressed a concern that there may be a limited window for settling Mars. Perhaps financial markets collapse. Perhaps there’s a worse pandemic. Perhaps a large asteroid hits the planet. Taking SpaceX public now is a bet that he can marshal the resources now, during his lifetime, to make Mars City One a reality. He is 54 years old.

The plan is not without risks, of course. If AI is something of a bubble, ten years from now, SpaceX may be sitting on hundreds of billions of dollars worth of satellites in space for which there is limited use. Maybe shareholders would rather SpaceX make them multimillionaires than make humans multiplanetary.

But Musk has never shied away from risks. So doubling down on his most successful asset in this moment is precisely what one would expect him to do.

Eric Berger is the senior space editor at Ars Technica, covering everything from astronomy to private space to NASA policy, and author of two books: Liftoff, about the rise of SpaceX; and Reentry, on the development of the Falcon 9 rocket and Dragon. A certified meteorologist, Eric lives in Houston.

After years of resisting it, SpaceX now plans to go public. Why? Read More »

A new open-weights AI coding model is closing in on proprietary options

agentic AI, AI, AI agents, AI coding, AI programming, Andrej Karpathy, ASML, Biz & IT, machine learning, Mistral AI, Open Source, open weights, software engineering, SWE-bench, vibe coding / Kris Guyer / December 10, 2025

On Tuesday, French AI startup Mistral AI released Devstral 2, a 123 billion parameter open-weights coding model designed to work as part of an autonomous software engineering agent. The model achieves a 72.2 percent score on SWE-bench Verified, a benchmark that attempts to test whether AI systems can solve real GitHub issues, putting it among the top-performing open-weights models.

Perhaps more notably, Mistral didn’t just release an AI model, it released a new development app called Mistral Vibe. It’s a command line interface (CLI) similar to Claude Code, OpenAI Codex, and Gemini CLI that lets developers interact with the Devstral models directly in their terminal. The tool can scan file structures and Git status to maintain context across an entire project, make changes across multiple files, and execute shell commands autonomously. Mistral released the CLI under the Apache 2.0 license.

It’s always wise to take AI benchmarks with a large grain of salt, but we’ve heard from employees of the big AI companies that they pay very close attention to how well models do on SWE-bench Verified, which presents AI models with 500 real software engineering problems pulled from GitHub issues in popular Python repositories. The AI must read the issue description, navigate the codebase, and generate a working patch that passes unit tests. While some AI researchers have noted that around 90 percent of the tasks in the benchmark test relatively simple bug fixes that experienced engineers could complete in under an hour, it’s one of the few standardized ways to compare coding models.

At the same time as the larger AI coding model, Mistral also released Devstral Small 2, a 24 billion parameter version that scores 68 percent on the same benchmark and can run locally on consumer hardware like a laptop with no Internet connection required. Both models support a 256,000 token context window, allowing them to process moderately large codebases (although whether you consider it large or small is very relative depending on overall project complexity). The company released Devstral 2 under a modified MIT license and Devstral Small 2 under the more permissive Apache 2.0 license.

A new open-weights AI coding model is closing in on proprietary options Read More »

Ugly infotainment mars the 2025 Subaru Forester Hybrid experience

2025 Subaru Forester Hybrid, car review, Cars / Kris Guyer / December 10, 2025

Although many of us associate it with rally-derived machinery from the late 1990s and early 2000s, these days, Subaru has mostly abandoned its performance cars to concentrate on its true calling—rugged, all-wheel-drive vehicles that are high on practicality, powered by horizontally opposed “boxer” engines. One area where the brand has never particularly excelled has been fuel efficiency, which is where today’s test car, the Subaru Forester Hybrid, comes in.

The last time Ars reviewed a Subaru Forester, it left us impressed. How about one with 40 percent better economy, in that case? Now, the 2.5 L flat-four engine operates on the Atkinson/Miller cycle, which generates 162 hp (121 kW) and 154 lb-ft (208 Nm). There’s an electric motor-generator starter and an electric traction motor with 118 hp (88 kW) and 199 lb-ft (270 Nm) that work together to send a combined 194 hp (145 kW) to all four wheels via a symmetrical all-wheel drive system and a planetary continuously variable transmission.

The Forester Hybrid is 183.3 inches (4,656 mm) long, 70.2 inches (1,783 mm) wide, and 68.1 inches (1,729 mm) tall, with a 105.1-inch (2,670 mm) wheelbase. Jonathan Gitlin

If that sounds vaguely familiar, that’s because it’s the same powertrain that Subaru has also fitted to the smaller Crosstrek Hybrid that we drove in September.

The 14 hp (10 kW) bump over the non-hybrid Forester is little enough that it probably won’t be noticed, but a combined EPA fuel efficiency of 35 mpg (6.7 L/100 km) is a meaningful increase over the unelectrified Forester’s 29 mpg (8.1 L/100 km). In practice, I struggled to exceed 31 mpg (7.6 L/100 km) during my week with the Forester, although as you’ll note from the temperatures displayed on the dash, winter temperatures have arrived, and we all know the cold makes all vehicles less efficient, not just EVs.

Driving a CVT hybrid can often feel disconcerting. Sometimes you press the accelerator and the car decides that electric propulsion will suffice; other times, the engine will rev with a non-linear relationship to the power being delivered at the wheels, as internal combustion tops up lithium-ion and then sends electrons to a motor to make the car go. Like all Atkinson cycle engines, it doesn’t sound very mellifluous when worked hard.

The rest of the driving experience was quite pleasant. The Forester Hybrid rides well on tires that have plenty of sidewall, and the 8.7-inches (221 mm) of ground clearance gives plenty of room for suspension travel. While it doesn’t handle like a WRX (or my dear departed Saab 9-2x), driving the Forester is no real chore. Other than the engine and some wind noise, it’s mostly refined on the move.

Ugly infotainment mars the 2025 Subaru Forester Hybrid experience Read More »

Author name: Kris Guyer