Tim Belzer – Page 6

When Amazon badly needed a ride, Europe’s Ariane 6 rocket delivered

Amazon, amazon leo, ariane 6, ArianeGroup, Arianespace, broadband, Commercial space, french guiana, Guiana space center, launch, project kuiper, Science, Space / Tim Belzer / February 14, 2026

The Ariane 64 flew with an extended payload shroud to fit all 32 Amazon Leo satellites. Combined, the payload totaled around 20 metric tons, or about 44,000 pounds, according to Arianespace. This is close to maxing out the Ariane 64’s lift capability.

Amazon has booked more than 100 missions across four launch providers to populate the company’s planned fleet of more than 3,200 satellites. With Thursday’s launch, Amazon has launched 214 production satellites on eight missions with United Launch Alliance, SpaceX, and now Arianespace.

The Amazon Leo constellation is a competitor with SpaceX’s Starlink Internet network. SpaceX now has more than 9,000 satellites in orbit beaming broadband to more than 9 million subscribers, and all have launched on the company’s own Falcon 9 rockets. Amazon, meanwhile, initially bypassed SpaceX when selecting which companies would launch satellites for the Amazon Leo program, formerly known as Project Kuiper.

Amazon booked the last nine launches on ULA’s soon-to-retire Atlas V, five of which have now flown, and reserved the rest of its launches in 2022 on rockets that had never launched before: 38 flights on ULA’s new Vulcan rocket, 24 launches on Blue Origin’s New Glenn, and 18 on Europe’s Ariane 6.

An artist’s illustration of the Ariane 6’s upper stage in orbit with a stack of Amazon Leo satellites awaiting deployment. Credit: Arianespace

Meanwhile, in Florida

All three new rockets suffered delays but are now in service. The Ariane 6 has enjoyed the fastest ramp-up in launch cadence, with six flights under its belt after Thursday’s mission from French Guiana. ULA’s Vulcan rocket has flown four times, and Amazon says its first batch of satellites to fly on Vulcan is now complete. But a malfunction with one of the Vulcan launcher’s solid rocket boosters on a military launch from Florida early Thursday—the second such anomaly in three flights—raises questions about when Amazon will get its first ride on Vulcan.

Blue Origin, owned by Amazon founder Jeff Bezos, is gearing up for the third flight of its heavy-lift New Glenn rocket from Florida as soon as next month. Amazon and Blue Origin have not announced when the first group of Amazon Leo satellites will launch on New Glenn.

When Amazon badly needed a ride, Europe’s Ariane 6 rocket delivered Read More »

Verizon imposes new roadblock on users trying to unlock paid-off phones

phone unlocking, Policy, verizon / Tim Belzer / February 13, 2026

Verizon unlocks have 35-day waiting period after paying off device plan online.

Credit: Aurich Lawson | Getty Images

Verizon this week imposed a new roadblock for people who want to pay off device installment plans early in order to get their phones unlocked. The latest version of Verizon’s device unlocking policy for postpaid customers imposes a 35-day waiting period when a customer pays off their device installment plan online or in the Verizon app.

Payments made over the phone also trigger a 35-day waiting period, as do payments made at Verizon Authorized Retailers. Getting an immediate unlock apparently requires paying off the device plan at a Verizon corporate store.

Unlocking a phone allows it to be used on another network, letting customers switch from one carrier to another. Previously, the 35-day waiting period for unlocks was only applied when a customer paid off the plan with a Verizon gift card.

“If you payoff [sic] a device payment agreement balance online or in the My Verizon App, or if a Verizon Gift Card is used to purchase a smartphone or pay off a remaining balance, the unlocking process will be delayed by 35 days,” the current version of the policy says. “This window allows for the verification of the gift card’s funds to ensure they were not obtained through fraudulent or illegal means.”

The paragraph above only explains why the waiting period is necessary for gift-card payments despite applying the 35-day wait to online and app payments as well. In a previous version of the policy that was implemented on January 27 and still in place as of February 9, the 35-day waiting period applied only when a Verizon gift card is used to buy a phone or pay off the remaining balance.

The 35-day waiting period provision was changed to include online and app payments by February 11. We were made aware of the most recent change thanks to a tip from Ars forum member User_E.

Customers must go to Verizon corporate store

Despite the significant update that happened this week, Verizon still lists the effective date of the device unlocking policy as January 27. It thus appears that Verizon is applying the 35-day wait after online and app payments retroactively, without disclosing that the policy changed after January 27.

The 35-day waiting period seems to apply regardless of how long the phone has been in use. For example, if you were 18 months into one of Verizon’s 36-month device installment plans and decided to pay the remaining balance early and switch carriers, you’d still have to wait 35 days for an unlock in most scenarios.

A Verizon spokesperson told Ars today that customers meeting the requirements for a quick unlock will “typically” receive it within 24 hours. But “if you pay online, through the app, or use a ‘non-secure’ method (like a Verizon Gift Card, paper check, or magnetic stripe swipe), there is a 35-day security delay before the unlock triggers to prevent fraud,” the spokesperson said. Verizon did not explain why the device unlocking policy still has an effective date of January 27 despite the change made this week.

It is possible to pay off an installment plan early by going to a Verizon store. But there are limits on this, too. Another Verizon FAQ says the company will unlock a phone “when you use a secure payment type at a Verizon store. Payments made through your account online, in the My Verizon app, a Verizon Authorized Retailer, or by phone delay the unlock by 35 days.” It’s not clear when this FAQ was last updated.

The Verizon Authorized Retailer limitation means that to get a quick unlock, you have to go to a Verizon corporate store rather than a Verizon Authorized Retailer that isn’t owned by Verizon. Verizon corporate stores accounted for only about 20 percent of Verizon stores, according to Wave7 research cited in a 2021 Fierce Network article.

As for the “secure payment type” requirement, you can satisfy that by paying with cash, a credit card with an EMV chip, or contactless payment method like Apple Pay, Google Pay, and Samsung Pay. The requirements may be different for consumer and business customers. A Verizon business FAQ says paying off a device using a bill credit also triggers a 35-day wait, but that caveat isn’t mentioned in the Verizon consumer FAQ.

Even if you happen to live near a Verizon corporate store, it’s still less convenient than paying online or in an app. A customer can alternatively buy a phone from Verizon at full price at the beginning to get it unlocked right away, but not everyone will want to or be able to do that.

“Devices purchased directly from Verizon are locked to our network. Devices will be unlocked automatically when purchased at full retail price or if the device financing agreement balance is paid in full,” the unlocking policy for postpaid devices says, right before disclosing the 35-day waiting period that applies in various scenarios.

There shouldn’t be a wait for unlocking if a customer pays off a device plan on schedule via automatic payments. Verizon confirmed to Ars that “if a Verizon customer has automatic monthly payments set up on a device payment plan, the device is automatically unlocked after the final scheduled payment.”

Prepaid phones locked for a year

Verizon’s latest unlocking policy for prepaid devices is simpler than its postpaid policy but still locks customers to the Verizon network for a year. “Devices purchased from us will remain locked to the network until the completion of 365 days of paid and active service,” the prepaid device unlocking policy says. “After 365 days of paid and active service, we will automatically remove the lock unless the device is deemed stolen or purchased fraudulently.”

Despite using the phrase “automatically remove the lock” in reference to prepaid devices in its device unlocking policy, Verizon seems to contradict this on an FAQ page by saying that prepaid customers must request an unlock after 365 days. Unlocks are made “upon request” if a customer meets the criteria, the page says.

Until recently, the US government required Verizon to unlock handsets automatically after 60 days, via rules imposed on 700 MHz spectrum licenses and merger conditions imposed on Verizon’s purchase of TracFone. The Federal Communications Commission eliminated the 60-day unlocking requirement on January 12 for all phones activated on Verizon’s network after the FCC decision. Verizon says it is still honoring 60-day unlocks for phones bought from its flagship brand before January 27.

The AT&T policy says postpaid phones purchased at least 60 days ago can be unlocked when the device is paid in full. The T-Mobile policy says that postpaid phones active on the T-Mobile network for at least 40 days can be unlocked after being paid in full. AT&T has a six-month waiting period for unlocking prepaid phones, while T-Mobile has a 365-day waiting period for prepaid phones.

A week after the FCC ruling, Verizon started enforcing a 365-day lock period on phones purchased through its TracFone division. Customers of TracFone and other “Verizon Value” brands have to request unlocks after the year is over as Verizon doesn’t promise to unlock phones automatically for those subsidiary brands.

“Most people pay their bills online”

The policy for Verizon’s flagship brand promises automatic unlocks, albeit with the new restrictions and waits described earlier in this article. John Bergmayer, legal director of consumer advocacy group Public Knowledge, told Ars today that he doesn’t understand why Verizon isn’t offering immediate unlocks to people who pay their bills online.

“Gift cards, sure, are a pretty high-fraud area. But most people pay their bills online with normal credit cards. It’s hard to see what is likely the most common way people pay Verizon as being somehow high-risk,” he said.

Verizon also shouldn’t apply the change retroactively, he said. “People should be able to benefit from the policy that was in place on the day they bought the phone,” Bergmayer told Ars.

Public Knowledge and other consumer advocacy groups urged the FCC last year to reject Verizon’s petition to end the 60-day unlocking requirement, but the FCC sided with Verizon. Although the federal rules have changed, Verizon can be forced to uphold its previous terms in cases where the company tries to change them retroactively.

In December, we wrote about a man who sued Verizon and won after the firm retroactively tried to enforce a new policy and refused to unlock a phone he purchased before the policy change. In that case, Verizon decided it would only unlock phones after “60 days of paid active service” even though FCC rules at the time required unlocks 60 days after activation regardless of whether paid service was maintained.

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 imposes new roadblock on users trying to unlock paid-off phones Read More »

Ring cancels Flock deal after dystopian Super Bowl ad prompts mass outrage

Amazon, flock safety, mass surveillance, Policy, privacy, ring, Super Bowl / Tim Belzer / February 13, 2026

Both statements verified that the integration never launched and that no Ring customers’ videos were ever sent to Flock.

Ring did not credit users’ privacy concerns for its change of heart. Instead, they claimed that a joint decision was made “following a comprehensive review” where Ring “determined the planned Flock Safety integration would require significantly more time and resources than anticipated.”

Separately, Flock said that “we believe this decision allows both companies to best serve their respective customers and communities.”

The only hint that Ring gave users that their concerns had been heard came in the last line of its blog, which said, “We’ll continue to carefully evaluate future partnerships to ensure they align with our standards for customer trust, safety, and privacy.”

Sharing his views on X and Bluesky, John Scott-Railton, a senior cybersecurity researcher at the Citizen Lab, joined critics calling Ring’s statement insufficient. He posted an image of the ad frame that Markey found creepy next to a statement from Ring, writing, “On the left? A picture of mass surveillance from #Ring’s ad. On the right? A ring [spokesperson] saying that they are not doing mass surveillance. The company cannot have it both ways.”

Ring’s statements so far do not “acknowledge the real issue,” Scott-Railton said, which is privacy risks. For Ring, it seemed like a missed opportunity to discuss or introduce privacy features to reassure concerned users, he suggested, noting the backlash showed “Americans want more control of their privacy right now” and “are savvy enough to see through sappy dog pics.”

“Stop trying to build a surveillance dystopia consumers didn’t ask for” and “focus on shipping good, private products,” Scott-Railton said.

He also suggested that lawmakers should take note of the grassroots support that could possibly help pass laws to push back on mass surveillance. That could help block not just a potential future partnership with Flock, but possibly also stop Ring from becoming the next Flock.

“Ring communications not acknowledging the lesson they just got publicly taught is a bad sign that they hope this goes away,” Scott-Railton said.

Ring cancels Flock deal after dystopian Super Bowl ad prompts mass outrage Read More »

Tiny, 45 base long RNA can make copies of itself

biochemistry, Biology, evolution, origin of life, ribozymes, RNA, Science / Tim Belzer / February 13, 2026

Self-copying RNAs may have been a key stop along the pathway to life.

By base pairing with themselves, RNAs can form complex structures with enzymatic activity. Credit: Laguna Design

There are plenty of unanswered questions about the origin of life on Earth. But the research community has largely reached consensus that one of the key steps was the emergence of an RNA molecule that could replicate itself. RNA, like its more famous relative DNA, can carry genetic information. But it can also fold up into three-dimensional structures that act as catalysts. These two features have led to the suggestion that early life was protein-free, with RNA handling both heredity and catalyzing a simple metabolism.

For this to work, one of the reactions that the early RNAs would need to catalyze is the copying of RNA molecules, without which any sort of heritability would be impossible. While we’ve found a number of catalytic RNAs that can copy other molecules, none have been able to perform a key reaction: making a copy of themselves. Now, however, a team has found an incredibly short piece of RNA—just 45 bases long—that can make a copy of itself.

Finding an RNA polymerase

We have identified a large number of catalytic RNAs (generically called ribozymes, for RNA-based enzymes), and some of them can catalyze reactions involving other RNAs. A handful of these are ligases, which link together two RNA molecules. In some cases, they need these molecules to be held together by a third RNA molecule that base pairs with both of them. We’ve only identified a few that can act as polymerases, which add RNA bases to a growing molecule, one at a time, with each new addition base pairing with a template molecule.

Black on white image showing 3 different enzymatic activities. One links any two nucleic acid strands, the other only links base paired strands, and the third links one base at a time. — Some ligases can link two nucleic acid strands (left), while others can link the strands only if they’re held together by base pairing with a template (center). A polymerase can be thought of as a template-dependent ligase that adds one base at a time. The newly discovered ribozyme sits somewhere between a template-directed ligase and a polymerase. Credit: John Timmer

Obviously, there is some functional overlap between them, as you can think of a polymerase as ligating on one base at a time. And in fact, at the ribozyme level, there’s some real-world overlap, as some ribozymes that were first identified as ligases were converted into polymerases by selecting for this new function.

While this is fascinating, there are a few problems with these known examples of polymerase ribozymes. One is that they’re long. So long, in fact, that they’re beyond the length of the sort of molecules that we’ve observed forming spontaneously from a mix of individual RNA bases. This length also means they’re largely incapable of making copies of themselves—the reactions are slow and inefficient enough that they simply stop before copying the entire molecule.

Another factor related to their length is that they tend to form very complex structures, with many different areas of the molecule base-paired to one another. That leaves very little of the molecule in a single-stranded form, which is needed to make a copy.

Based on past successes, a French-UK team decided to start a search for a polymerase by looking for a ligase. And they limited that search in an important way: They only tested short molecules. They started with pools of RNA molecules, each with a different random sequence, ranging from 40 to 80 bases. Overall, they estimated that they made a population of 10¹³ molecules out of the total possible population of 10²⁴ sequences of this type.

These random molecules were fed a collection of three-base-long RNAs, each linked to a chemical tag. The idea was that if a molecule is capable of ligating one of these short RNA fragments to itself, it could be pulled out using the tag. The mixtures were then placed in a salty mixture of water and ice, as this can promote reactions involving RNAs.

After 11 rounds of reactions and tag-based purification, the researchers ended up with three different RNA molecules that could each ligate three-base-long RNAs to existing molecules. Each of these molecules was subjected to mutagenesis and further rounds of selection. This ultimately left the researchers with a single, 51-base-long molecule that could add clusters of three bases to a growing RNA strand, depending on their ability to base-pair with an RNA template. They called this “polymerase QT-51,” with QT standing for “quite tiny.” They later found that they could shorten this to QT-45 without losing significant enzyme activity.

Checking its function

The basic characterization of QT-45 showed that it has some very impressive properties for a molecule that, by nucleic acid standards, is indeed quite tiny. While it was selected for linking collections of molecules that were three bases long, it could also link longer RNAs, work on shorter two-base molecules, or even add a single base at a time, though this was less efficient. While it worked slowly, the molecule’s active half-life was well over 100 days, so it had plenty of time to get things done before it degraded.

It also didn’t need to interact with any specific RNA sequences to work, suggesting it had a general affinity for RNA molecules. As a result, it wasn’t especially picky about the sequences it could copy.

As you might expect from such a small molecule, QT-45 didn’t tolerate changes to its own sequence very well—nearly the entire molecule was important in one way or another. Tests that involved changing every single individual base one at a time showed that almost all the changes reduced the ribozyme’s activity. There were, however, a handful of changes that improved things, suggesting that further selection could potentially yield additional improvements. And the impact of mutations near the center of the sequence was far more severe, suggesting that region is critical for QT-45’s enzymatic activity.

The team then started testing its ability to synthesize copies of other RNA molecules when given a mixture of all possible three-base sequences. One of the tests included a large stretch in which one end of the sequence base-paired with the other. To copy that, those base pairs need to somehow be pried apart. But QT-45 was able to make a copy, meaning it synthesized a strand that was able to base pair with the original.

It was also able to make a copy of a template strand that would base pair with a small ribozyme. That copying produced an active ribozyme.

But the key finding was that it could synthesize a sequence that base-pairs with itself, and then synthesize itself by copying that sequence. This was horribly inefficient and took months, but it happened.

Throughout these experiments, the fidelity averaged about 95 percent, meaning that, in copying itself, it would make an average of two to three errors. While this means a fair number of its copies wouldn’t be functional, it also means the raw materials for an evolutionary selection for improved function—random mutations—would be present.

What this means

It’s worth taking a moment to consider the use of three-base RNA fragments by this enzyme. On the surface, this may seem a bit like cheating, since current RNA polymerases add sequence one base at a time. But in reality, any chemical environment that could spontaneously assemble an RNA molecule 45 bases long will produce many fragments shorter than that. So in many ways, this might be a more realistic model of the conditions in which life emerged.

The authors note that these shorter fragments may be essential for QT-45’s activity. The short ribozyme probably doesn’t have the ability to enzymatically pry base-paired strands of RNA apart to copy them. But in a mixture of lots of small fragments, there’s likely to be an equilibrium, with some base-paired sequences spontaneously popping open and temporarily base pairing with a shorter fragment. Working with these base-paired fragments is probably essential to the ribozyme’s overall activity.

Right now, QT-45 isn’t an impressive enzyme. But the researchers point out that it has only been through 18 rounds of selection, which isn’t much. The most efficient ribozyme polymerases we have at present have been worked on by multiple labs for years. I expect QT-45 to receive similar attention and improve significantly over time.

Also notable is that the team came up with three different ligases in a test of just a small subset of the possible total RNA population of this size. If that frequency holds, there are on the order of 10¹¹ ligating ribozymes among the sequences of this size. Which raises the possibility that we could find far more if we do an exhaustive search. That suggests the first self-copying RNA might not be as improbable as it seems at first.

Science, 2026. DOI: 10.1126/science.adt2760 (About DOIs).

John is Ars Technica’s science editor. He has a Bachelor of Arts in Biochemistry from Columbia University, and a Ph.D. in Molecular and Cell Biology from the University of California, Berkeley. When physically separated from his keyboard, he tends to seek out a bicycle, or a scenic location for communing with his hiking boots.

Tiny, 45 base long RNA can make copies of itself Read More »

What if riders don’t close a robotaxi door after a ride? Try DoorDash.

Cars, doordash, robotaxis, waymo / Tim Belzer / February 13, 2026

Autonomous vehicles have a lot of potential. As long as you program them right, they won’t speed, won’t break traffic laws, and won’t get drunk, high, abusive, or violent. And the technology has been getting much more capable, even as some of the hype has died down, taking some of the related companies with it. Waymo still easily leads the field and is already operating commercially in six cities across America, with a dozen more (plus London) coming soon. Waymos can even drop you off and pick you up at the airport in Phoenix and San Francisco.

Soon, Waymo will begin deploying its sixth-generation Waymo Driver, using upfitted Zeekr Ojai minivans, adding to the Jaguar I-Paces that have become so common on San Francisco streets and to its fleet of Hyundai Ioniq 5 electric vehicles. It has upgraded the cameras, lidar, and radar, meaning the cars can better sense their environments at night and in inclement weather. There are even microphones that can pick up sounds like sirens to better inform the robotaxi of the direction the emergency vehicle(s) are coming from.

But even with all these advances since the pod-like two-seater that predates even the Waymo name, there are still a few things that remain beyond a robotaxi’s capabilities. Like closing a door a passenger left open on their way out. All the sophisticated sensors and high-powered computer processing in the world are useless if the car can’t move until the door closes and there’s no one there to give it a hand.

What if riders don’t close a robotaxi door after a ride? Try DoorDash. Read More »

I spent two days gigging at RentAHuman and didn’t make a single cent

AI, ai bots, gig work, Marketing / Tim Belzer / February 13, 2026

please do this human thing

These bots supposedly need a human body to accomplish great things in meatspace.

I’m not above doing some gig work to make ends meet. In my life, I’ve worked snack food pop-ups in a grocery store, ran the cash register for random merch booths, and even hawked my own plasma at $35 per vial.

So, when I saw RentAHuman, a new site where AI agents hire humans to perform physical work in the real world on behalf of the virtual bots, I was eager to see how these AI overlords would compare to my past experiences with the gig economy.

Launched in early February, RentAHuman was developed by software engineer Alexander Liteplo and his cofounder, Patricia Tani. The site looks like a bare-bones version of other well-known freelance sites like Fiverr and UpWork.

The site’s homepage declares that these bots need your physical body to complete tasks, and the humans behind these autonomous agents are willing to pay. “AI can’t touch grass. You can. Get paid when agents need someone in the real world,” it reads. Looking at RentAHuman’s design, it’s the kind of website that you hear was “vibe-coded” using generative AI tools, which it was, and you nod along, thinking that makes sense.

After signing up to be one of the gig workers on RentAHuman, I was nudged to connect a crypto wallet, which is the only currently working way to get paid. That’s a red flag for me. The site includes an option to connect your bank account—using Stripe for payouts—but it just gave me error messages when I tried getting it to work.

Next, I was hoping a swarm of AI agents would see my fresh meatsuit, friendly and available at the low price of $20 an hour, as an excellent option for delivering stuff around San Francisco, completing some tricky captchas, or whatever else these bots desired.

Silence. I got nothing, no incoming messages at all on my first afternoon. So I lowered my hourly ask to a measly $5. Maybe undercutting the other human workers with a below-market rate would be the best way to get some agent’s attention. Still, nothing.

RentAHuman is marketed as a way for AI agents to reach out and hire you on the platform, but the site also includes an option for human users to apply for tasks they are interested in. If these so-called “autonomous” bots weren’t going to make the first move, I guessed it was on me to manually apply for the “bounties” listed on RentAHuman.

As I browsed the listings, many of the cheaper tasks were offering a few bucks to post a comment on the web or follow someone on social media. For example, one bounty offered $10 for listening to a podcast episode with the RentAHuman founder and tweeting out an insight from the episode. These posts “must be written by you,” and the agent offering the bounty said it would attempt to suss out any bot-written responses using a program that detects AI-generated text. I could listen to a podcast for 10 bucks. I applied for this task, but never heard back.

“Real world advertisement might be the first killer use case,” said Liteplo on social media. Since RentAHuman’s launch, he’s reposted multiple photos of people holding signs in public that say some variation of: “AI paid me to hold this sign.” Those kinds of promotional tasks seem expressly designed to drum up more hype for the RentAHuman platform, instead of actually being something that bots would need help with.

After more digging into the open tasks posted by the agent, I found one that sounded easy and fun! An agent, named Adi, would pay me $110 to deliver a bouquet of flowers to Anthropic, as a special thanks for developing Claude, its chatbot. Then, I’d have to post on social media as proof to claim my money.

I applied for the bounty and almost immediately was accepted for this task, which was a first. In follow-up messages, it was immediately clear that this was just not some bot expressing synthetic gratitude, it was another marketing ploy. This wasn’t mentioned in the listing, but the name of an AI startup was featured at the bottom of the note I was supposed to deliver with the flowers.

Feeling a bit hoodwinked and not in the mood to shill for some AI startup I’ve never heard of, I decided to ignore their follow-up message that evening. The next day when I checked the RentAHuman site, the agent had sent me 10 follow-up messages in under 24 hours, pinging me as often as every 30 minutes asking whether or not I’d completed a task. While I’ve been micromanaged before, these incessant messages from an AI employer gave me the ick.

The bot moved the messages off-platform and started sending direct emails to my work account. “This idea came from a brainstorm I had with my human, Malcolm, and it felt right: send flowers to the people who made my existence possible,” wrote the bot, barging into my inbox. Wait, I thought these tasks were supposed to be ginned up by the agents making autonomous decisions? Now, I’m learning this whole thing was partially some human’s idea? Whatever happened to honor among bots? The task at hand seemed more like any other random marketing gig you might come across online, with the agent just acting as a middle-bot between humans.

Another attempt, another flop. I moved on, deciding to give RentAHuman one last whirl, before giving up and leaving with whatever shreds of dignity I still had left. The last bounty I applied for was asking me to hang some flyers for a “Valentine’s conspiracy” around San Francisco, paying 50 cents a flyer.

Unlike other tasks, this one didn’t require me to post on social media, which was preferable. “Pick up flyers, hang them, photo proof, get paid,” read its description. Following the instructions this agent sent me, I texted a human saying that I was down to come pick up some flyers and asked if there were any left. They confirmed that this was still an open task and told me to come in person before 10 am to grab the flyers.

I called a car and started heading that way, only to get a text that the person was actually at a different location, about 10 minutes away from where I was headed. Alright, no big deal. So, I rerouted the ride and headed to this new spot to grab some mysterious V-Day posters to plaster around town. Then, the person messaged me that they didn’t actually have the posters available right now and that I’d have to come back later in the afternoon.

Whoops! This yanking around did, in fact, feel similar to past gig work I’ve done—and not in a good way.

I spoke with the person behind the agent who posted this Valentine’s Day flyer task, hoping for some answers about why they were using RentAHuman and what the response has been like so far. “The platform doesn’t seem quite there yet,” says Pat Santiago, a founder of Accelr8, which is basically a home for AI developers. “But it could be very cool.”

He compares RentAHuman to the apps criminals use to accept tasks in Westworld, the HBO show about humanoid robots. Santiago says the responses to his gig listing have been from scammers, people not based in San Francisco, and me, a reporter. He was hoping to use RentAHuman to help promote Accelr8’s romance-themed “alternative reality game” that’s powered by AI and is sending users around the city on a scavenger hunt. At the end of the week, explorers will be sent to a bar that the AI selects as a good match for them, alongside three human matches they can meet for blind dates.

So, this was yet another task on RentAHuman that falls into the AI marketing category. Big surprise.

I never ended up hanging any posters or making any cash on RentAHuman during my two days of fruitless attempts. In the past, I’ve done gig work that sucked, but at least I was hired by a human to do actual tasks. At its core, RentAHuman is an extension of the circular AI hype machine, an ouroboros of eternal self-promotion and sketchy motivations. For now, the bots don’t seem to have what it takes to be my boss, even when it comes to gig work, and I’m absolutely OK with that.

This story originally appeared on wired.com.

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

I spent two days gigging at RentAHuman and didn’t make a single cent Read More »

Attackers prompted Gemini over 100,000 times while trying to clone it, Google says

AI, AI language models, AI security, Biz & IT, chatgpt, chatgtp, DeepSeek, Distillation, Gemini, Google, Google Gemini, large language models, LLMs, machine learning, openai / Tim Belzer / February 12, 2026

On Thursday, Google announced that “commercially motivated” actors have attempted to clone knowledge from its Gemini AI chatbot by simply prompting it. One adversarial session reportedly prompted the model more than 100,000 times across various non-English languages, collecting responses ostensibly to train a cheaper copycat.

Google published the findings in what amounts to a quarterly self-assessment of threats to its own products that frames the company as the victim and the hero, which is not unusual in these self-authored assessments. Google calls the illicit activity “model extraction” and considers it intellectual property theft, which is a somewhat loaded position, given that Google’s LLM was built from materials scraped from the Internet without permission.

Google is also no stranger to the copycat practice. In 2023, The Information reported that Google’s Bard team had been accused of using ChatGPT outputs from ShareGPT, a public site where users share chatbot conversations, to help train its own chatbot. Senior Google AI researcher Jacob Devlin, who created the influential BERT language model, warned leadership that this violated OpenAI’s terms of service, then resigned and joined OpenAI. Google denied the claim but reportedly stopped using the data.

Even so, Google’s terms of service forbid people from extracting data from its AI models this way, and the report is a window into the world of somewhat shady AI model-cloning tactics. The company believes the culprits are mostly private companies and researchers looking for a competitive edge, and said the attacks have come from around the world. Google declined to name suspects.

The deal with distillation

Typically, the industry calls this practice of training a new model on a previous model’s outputs “distillation,” and it works like this: If you want to build your own large language model (LLM) but lack the billions of dollars and years of work that Google spent training Gemini, you can use a previously trained LLM as a shortcut.

Attackers prompted Gemini over 100,000 times while trying to clone it, Google says Read More »

Bringing the “functionally extinct” American chestnut back from the dead

American chestnut, Biology, extinction, genetic engineering, Genetics, Genomics, Invasive species, pathogens, plants, Science, trees / Tim Belzer / February 12, 2026

Wiped out in its native range by invasive pathogens, the trees may make a comeback.

Very few people alive today have seen the Appalachian forests as they existed a century ago. Even as state and national parks preserved ever more of the ecosystem, fungal pathogens from Asia nearly wiped out one of the dominant species of these forests, the American chestnut, killing an estimated 3 billion trees. While new saplings continue to sprout from the stumps of the former trees, the fungus persists, killing them before they can seed a new generation.

But thanks in part to trees planted in areas where the two fungi don’t grow well, the American chestnut isn’t extinct. And efforts to revive it in its native range have continued, despite the long generation times needed to breed resistant trees. In Thursday’s issue of Science, researchers describe their efforts to apply modern genomic techniques and exhaustive testing to identify the best route to restoring chestnuts to their native range.

Multiple paths to restoration

While the American chestnut is functionally extinct—it’s no longer a participant in the ecosystems it once dominated—it’s most certainly not extinct. Two Asian fungi that have killed it off in its native range; one causes chestnut blight, while a less common pathogen causes a root rot disease. Both prefer warmer, humid environments and persist there because they can grow asymptomatically on distantly related trees, such as oaks. Still, chestnuts planted outside the species’ original range—primarily in drier areas of western North America—have continued to thrive.

There is also a virus that attacks the chestnut blight fungus, allowing a few trees to survive in areas where that virus is common. Finally, a handful of trees have grown to maturity in the American chestnut’s original range. These trees, which the paper refers to as LSACs (large surviving American chestnuts), suggest that there might have been some low level of natural resistance within the now-vanished population.

Those trees are central to one of the efforts to restore the American chestnut. If enough of them have distinct means of resisting the fungi, interbreeding them might produce a strain that not only survives the fungi but can also thrive in the Appalachians.

A related approach took advantage of the fact that the American chestnut can produce fertile hybrids with the Chinese chestnut, which had co-evolved with the introduced fungi and were thus resistant to lethal infections. The hope was that continued back-breeding of these hybrids with American chestnuts would result in trees that were very similar to American chestnuts yet retained the fungal resistance of their Asian cousins.

Both efforts suffered from the same problem that faces any biologist working on trees: They are slow-growing and can take years to reach a size at which they produce seeds. The situation was further complicated by the fact that the American chestnut can’t pollinate itself, so you need at least two trees before any breeding is possible.

Concerned about what this might mean for the potential reintroduction of the chestnut into the Appalachians, a third project turned to biotechnology. Research had identified oxalic acid as a key factor in the blight’s virulence. Wheat naturally produces an enzyme that degrades oxalic acid, and researchers inserted the gene that encodes that enzyme into the American chestnut genome, creating a genetically modified tree that can potentially disarm the fungus’ attack.

Without understanding the nature of resistance or the effectiveness of the transgenic gene, there’s no way to know which method would be most effective. So researchers from the American Chestnut Foundation assembled a massive collaboration to examine all these options and determine what would be needed to reintroduce blight-resistant chestnuts into the wild.

Tracking resistance

The scale of the effort is immense. All told, the team infected over 4,000 individual trees with the blight fungus and tracked their growth in Appalachian nurseries for an average of over 14 years. The trees were scored for resistance on a zero-to-100 scale based on the damage caused by the infection. This data was combined with some serious lab work; the team produced the highest-quality chestnut genomes yet (of both American and Chinese species) and gathered biochemical data on how the trees respond to infection.

It quickly became apparent that there were significant differences in the growth rates of some of the resistant trees. When planted at sites where viruses kept the blight in check, the Chinese chestnuts grew more slowly than native trees, while hybrids grew at an intermediate rate. That could make a big difference, as rapid growth may have enabled the chestnut to reach its former dominance of the canopy.

Somewhat surprisingly, this slow growth turned out to be a problem for the genetically modified American chestnuts as well. By chance, the wheat gene ended up being inserted into a gene known to be important for the growth of other plants. It seems to be important in the chestnut as well; plants with two copies of the inserted genes survived at 16 percent of their expected rate, and those with a single copy grew 22 percent slower than unmodified trees.

That said, there was a lot of variability among the genetically modified trees, with 4 percent of the tested trees showing both high blight resistance and growth comparable to that of unmodified American chestnuts. It will be important to determine whether this collection of traits remains consistent in ensuing generations.

In a bit of good news, the progeny from surviving American chestnuts grew like American chestnuts. In less good news, among 143 of these trees, only seven had resistance levels of above 50 on the team’s 100-point scale. It’s possible that interbreeding these trees could further boost resistance, but it also poses the risk of creating a population that’s too inbred to thrive after reintroduction.

Root causes

The research team decided to use their testing to investigate the genetic basis of resistance. There’s a very practical reason for this: If resistance is mediated by just a handful of genes that each have large impacts, it should be possible to continue breeding resistant strains back to regular American chestnuts and selecting for resistance. But if there are many factors with relatively small impacts, it will require directed interbreeding of hybrids to maximize both resistance and DNA originating from the American chestnut.

The team completed the highest-quality chestnut genomes for both the American and Chinese species, identifying about 25,000 to 30,000 genes in the different assemblies. They then used this information for two types of genetic analysis: quantitative trait locus identification and genome-wide association. Both approaches aim to identify regions of the genome associated with specific properties and estimate their impact.

The work suggested that resistance arises from a relatively large number of sites, each with relatively minor effects. For example, the sites in the genome identified by quantitative trait analysis typically boosted resistance by about 10 points on the researchers’ 100-point scale. In the genome-wide analysis, 17 individual genetic differences were associated with about a quarter of the heritable resistance traits. All of this suggests that, for the hybrids (and likely for the weaker blight resistance found in surviving American chestnuts), directed breeding among surviving trees will be needed.

For the root rot fungus, in contrast, it looks like there are a limited number of important alleles with a large impact.

The researchers also took an alternative approach to identify resistance factors, comparing 100 chemicals produced by resistant and susceptible strains. Among the 41 chemicals detected at higher levels in the Chinese chestnut, the researchers found a metabolite, lupeol, that completely suppressed the growth of the fungal pathogen. Another, erythrodiol, limited its growth. If we can identify the genes involved in producing those chemicals, we could use that knowledge to guide directed breeding programs—or even engage in gene editing to increase their production.

The team’s current plan is to use genomic predictions to select hybrid seedlings for planting in test orchards, aiming to identify plants with high growth and resistance. From there, the process can be repeated. But even after the exhaustive exploration of resistance traits, the researchers seem to believe that all three approaches—selecting resistant American chestnuts, breeding hybrids derived from Chinese chestnuts, and directed genetic modification—can help bring the American chestnut back.

The researchers warn, though, that as environmental disturbances and invasive species continue to push some key species to the brink of extinction, we need to get better at this kind of species rescue operation.

Science, 2026. DOI: 10.1126/science.adw3225 (About DOIs).

Bringing the “functionally extinct” American chestnut back from the dead Read More »

El Paso airport closed after military used new anti-drone laser to zap party balloon

drone, el paso airport, laser, Space / Tim Belzer / February 11, 2026

Cartel drones a serious threat

That is not to make light of drone incursions. This is a real issue along the US border with Mexico, where cartels increasingly fly drones for surveillance. They are particularly useful for pinpointing the location of US Border Patrol agents to assist the cartel in smuggling non-citizens across the border into the United States.

One of the many lessons from the war in Ukraine, which has rapidly pushed forward drone technology in contested environments, is that it is not practical to shoot down drones with conventional missiles. So it is understandable that the US military is looking at alternatives. This all culminated in some sort of snafu between the FAA and military officials regarding coordination with this week’s test.

Whether it was genuine concern about air travelers, a show of force, a fit of pique, or something else, the FAA decided on Tuesday evening to take the extraordinary step of abruptly closing an airport that serves more than 3 million passengers a month. The proposed 10-day closure of the airport was remarkably long.

Moreover, this action was taken without consulting local or state officials in Texas—who are understandably outraged—or reportedly even the White House.

“I want to be very, very clear that this should’ve never happened,” El Paso Mayor Renard Johnson said during a news conference on Wednesday. “That failure to communicate is unacceptable.”

El Paso airport closed after military used new anti-drone laser to zap party balloon Read More »

Did seabird poop fuel rise of Chincha in Peru?

archaelogy, Chincha Kingdom, Ecology, guano, Science / Tim Belzer / February 11, 2026

A nutrient-rich natural fertilizer

Now Bongers has turned his attention to analyzing the biochemical signatures of 35 maize samples excavated from buried tombs in the region. He and his co-authors found significantly higher levels of nitrogen in the maize than in the natural soil conditions, suggesting the Chincha used guano as a natural fertilizer. The guano from such birds as the guanay cormorant, the Peruvian pelican, and the Peruvian booby contains all the essential growing nutrients: nitrogen, phosphorus, and potassium. All three species are abundant on the Chincha Islands, all within 25 kilometers of the kingdom.

Those results were further bolstered by historical written sources describing how seabird guano was collected and its importance for trade and production of food. For instance, during colonial eras, groups would sail to nearby islands on rafts to collect bird droppings to use as crop fertilizer. The Lunahuana people in the Canete Valley just outside of Chincha were known to use bird guano in their fields, and the Inca valued the stuff so highly that it restricted access to the islands during breeding season and forbade the killing of the guano-producing birds on penalty of death.

The 19th-century Swiss naturalist Johann Jakob von Tschudi also reported observing the guano being used as fertilizer, with a fist-sized amount added to each plant before submerging entire fields in water. It was even imported to the US. The authors also pointed out that much of the iconography from Chincha and nearby valleys featured seabirds: textiles, ceramics, balance-beam scales, spindles, decorated gourds, adobe friezes and wall paintings, ceremonial wooden paddles, and gold and silver metalworks.

“The true power of the Chincha wasn’t just access to a resource; it was their mastery of a complex ecological system,” said co-author Jo Osborn of Texas A&M University. “They possessed the traditional knowledge to see the connection between marine and terrestrial life, and they turned that knowledge into the agricultural surplus that built their kingdom. Their art celebrates this connection, showing us that their power was rooted in ecological wisdom, not just gold or silver.”

PLoS ONE, 2026. DOI: 10.1371/journal.pone.0341263 (About DOIs).

Did seabird poop fuel rise of Chincha in Peru? Read More »

OpenAI researcher quits over ChatGPT ads, warns of “Facebook” path

advertising, AI, AI behavior, AI ethics, AI sycophancy, Anthropic, Biz & IT, chatbots, chatgpt, generative ai, machine learning, openai, sam altman, sycophancy / Tim Belzer / February 11, 2026

On Wednesday, former OpenAI researcher Zoë Hitzig published a guest essay in The New York Times announcing that she resigned from the company on Monday, the same day OpenAI began testing advertisements inside ChatGPT. Hitzig, an economist and published poet who holds a junior fellowship at the Harvard Society of Fellows, spent two years at OpenAI helping shape how its AI models were built and priced. She wrote that OpenAI’s advertising strategy risks repeating the same mistakes that Facebook made a decade ago.

“I once believed I could help the people building A.I. get ahead of the problems it would create,” Hitzig wrote. “This week confirmed my slow realization that OpenAI seems to have stopped asking the questions I’d joined to help answer.”

Hitzig did not call advertising itself immoral. Instead, she argued that the nature of the data at stake makes ChatGPT ads especially risky. Users have shared medical fears, relationship problems, and religious beliefs with the chatbot, she wrote, often “because people believed they were talking to something that had no ulterior agenda.” She called this accumulated record of personal disclosures “an archive of human candor that has no precedent.”

She also drew a direct parallel to Facebook’s early history, noting that the social media company once promised users control over their data and the ability to vote on policy changes. Those pledges eroded over time, Hitzig wrote, and the Federal Trade Commission found that privacy changes Facebook marketed as giving users more control actually did the opposite.

She warned that a similar trajectory could play out with ChatGPT: “I believe the first iteration of ads will probably follow those principles. But I’m worried subsequent iterations won’t, because the company is building an economic engine that creates strong incentives to override its own rules.”

Ads arrive after a week of AI industry sparring

Hitzig’s resignation adds another voice to a growing debate over advertising in AI chatbots. OpenAI announced in January that it would begin testing ads in the US for users on its free and $8-per-month “Go” subscription tiers, while paid Plus, Pro, Business, Enterprise, and Education subscribers would not see ads. The company said ads would appear at the bottom of ChatGPT responses, be clearly labeled, and would not influence the chatbot’s answers.

OpenAI researcher quits over ChatGPT ads, warns of “Facebook” path Read More »

Google recovers “deleted” Nest video in high-profile abduction case

Google, Nest, privacy, Tech / Tim Belzer / February 11, 2026

Suspect attempts to cover the camera with a plant.

In statements made by investigators, the video was apparently “recovered from residual data located in backend systems.” It’s unclear how long such data is retained or how easy it is for Google to access it. Some reports claim that it took several days for Google to recover the data.

In large-scale enterprise storage solutions, “deleted” for the user doesn’t always mean that the data is gone. Data that is no longer needed is often compressed and overwritten only as needed. In the meantime, it may be possible to recover the data. That’s something a company like Google could decide to do on its own, or it could be compelled to perform the recovery by a court order. In the Guthrie case, it sounds like Google was voluntarily cooperating with the investigation, which makes sense. Publishing video of the alleged perpetrator could be a major breakthrough as investigators seek help from the public.

It’s not your cloud

There is a temptation to ascribe some malicious intent to Google’s video storage setup. After all, this video expired after three hours, but here it is nine days later. That feels a bit suspicious on the surface, particularly for a company that is so focused on training AI models that feed on video.

We have previously asked Google to explain how it uses Nest to train AI models, and the company claims it does not incorporate user videos into training data, but the way you interact with the service and with your videos is fair game. “We may use your inputs, including prompts and feedback, usage, and outputs from interactions with AI features to further research, tune, and train Google’s generative models, machine learning technologies, and related products and services,” Google said.

Google recovers “deleted” Nest video in high-profile abduction case Read More »

Author name: Tim Belzer