Tech
You Can Disable Gemini in Chrome if It’s Freaking You Out
If you use Google’s Chrome browser for desktop, there’s probably a Gemini Nano AI model running on your computer right now and taking up about 4 GB of space. That’s not necessarily a bad thing, but if you didn’t know about it and don’t want it, there’s a way to turn it off.
The file started auto-downloading for Chrome users in 2024 after Google built Gemini Nano into the browser. But a report by That Privacy Guy this week and the ensuing reception it received highlighted how unaware many users were—perhaps a result of a flood of AI services and features across the tech industry that have been difficult for users to keep up with.
To uninstall the Gemini Nano file, open Chrome on your computer, in the top right corner click the “More” menu represented by three vertical dots, then go to Settings, System, and then toggle “On-device AI” to be off. The Privacy Guy article noted that if you directly uninstall the Gemini Nano file in the directory, Chrome will silently, automatically redownload it the next time the browser reboots.
A Google spokesperson tells WIRED that the company started rolling out the On-device AI toggle in February so users can turn off the features if they choose and remove the model. “Once disabled, the model will no longer download or update,” the spokesperson says in a statement. The company added, too, that the system is designed so Gemini Nano “will automatically uninstall if the device is low on resources.”
Google built the model into Chrome to enabled on-device AI scam-detection features. It was also aimed at providing a way for developers to integrate AI-related application programming interfaces while keeping data on users’ devices when possible and out of the cloud. These features are separate from Chrome’s AI Mode, which does not use the local Gemini Nano model.
Parisa Tabriz, Chrome’s general manager, emphasized in a post on X on Wednesday that integrating Gemini Nano “powers important security capabilities like on-device scam detection and developer APIs without sending your data to the cloud.”
Google certainly did announce the Gemini Nano integration into Chrome and discussed it publicly, but for users who simply use Chrome because it is the world’s biggest, most recognizable browser and don’t necessarily follow every granular update, the lack of an in-your-face notification about a large AI model file sitting and running on your computer may be upsetting.
Longtime security and compliance consultant Davi Ottenheimer says that he follows Chrome updates closely but could have easily missed the Gemini Nano integration. “An on-device model could be a hidden minefield,” he says. And the fact that Google launched the integration in 2024 but didn’t start rolling out a settings control for users to turn it off until February shows that, at least initially, the feature wasn’t conceived as something that users would interact with.
Just because you can remove Gemini Nano from Chrome doesn’t mean you necessarily should—or that doing so is better for your privacy.
Local processing is a more private way to utilize AI capabilities. If you remove the model, the features Google uses it for—including the AI-enabled scam detection—will cease to function. But since Gemini Nano is also used by Chrome to enable local AI processing for third-party developers, blocking this route could have a range of outcomes when interacting with non-Google web services in the browser. A Google spokesperson tells WIRED that if you turn off On-device AI, “certain security features will not be available, and sites that use the on device APIs will behave differently.”
Of course, if neither option seems right, there’s always an alternative: Use a different browser.
Tech
Trump Pivots on AI Regulation, Worker Ousted by DOGE Runs for Office, and Hantavirus Explained
Brian Barrett: This is the first time I’ve thought about contact tracing in many years, and I was so happy not thinking about it for so long, because it is such a complicated process and something that is really hard work to do. Emily, given all of that, what is the level of concern here, given what the World Health Organization has said and other organizations? It sounds like cautious about it, but maybe not freak out time yet, but I defer to you because maybe that’s just me trying to make myself feel better.
Emily Mullin: No, I think you’re right. The hantavirus expert I spoke with said there have been past clusters of the Andes strain before, but not big outbreaks. And these clusters have tended to involve prolonged close contact with people suffering from the disease. This is a virus that does not spread nearly as efficiently as other respiratory viruses that we’re used to like Covid or flu, for instance. Hantavirus symptoms are also typically pretty severe. So this is not a virus, again, like Covid where lots of people are going around infected with the disease, spreading it asymptomatically without knowing about it. So that’s at least a little bit of comfort, even though the flip side of that is that the disease is quite severe. So the World Health Organization says the risk to the general public is currently low, and this is probably not another Covid situation.
Brian Barrett: Leah, how we feeling?
Leah Feiger: Not good, you guys. I don’t know. Are you kidding? How are you feeling? Maybe this is my moment to go, “Are you with me yet?”
Brian Barrett: No, I was good, but then Emily hit that probably pretty hard in a way that I suddenly felt a little more anxious.
Leah Feiger: Yeah, it was the swallowing of the probably.
Emily Mullin: That was me editorializing. The World Health Organization did not include the probably.
Brian Barrett: OK. What if they had it just in italics or big quotation marks? Like it’s “probably” fine.
Leah Feiger: I don’t know, guys. I think, one, I’m fascinated that there’s different strains of this. And it brought me back so early on to the armchair scientists in early Covid who were like, “No, no, no, this is totally fine.” So for there to officially be announced, yes, this is the strain that can get passed between humans, I think is notable at the very least. Got to give me that.
Brian Barrett: Oh, I think that’s true. And I think my open questions are, how long do these people have to stay on this ship before everyone says, “OK, you can go now,” or do they send them back to shore and just have them isolate for a certain amount of time? The contact tracing is concerning because again, I’m having flashbacks. But I do think the things that, Emily, that you said about how this is different from Covid in important ways in terms of how quickly it can spread, how easily it can spread, especially now that we have the mechanisms in place to do these contact tracing things, I’m going to remain on my not too worried yet.
Tech
Here’s What You Need to Know About the Hantavirus
Cruises are so closely associated with illness that the highly contagious norovirus is commonly called the “cruise ship virus.”
But a ship headed for Spain’s Canary Islands has attracted global attention due to a rare outbreak of hantavirus that’s left three dead. While alarming, health officials and infectious disease experts say the risk to the general public right now is low because hantavirus is less contagious than other respiratory diseases like the coronavirus responsible for the Covid-19 pandemic.
“This is not Covid, this is not influenza. It spreads very, very differently,” Maria Van Kerkhove, director of epidemic and pandemic preparedness and prevention at the World Health Organization, said at a press conference on Thursday.
During the briefing, WHO director-general Tedros Adhanom Ghebreyesus confirmed eight hantavirus cases among passengers of the MV Hondius luxury cruise ship, including the three who died. Typically transmitted by rodents, hantavirus can cause severe disease in humans. People usually get sick by inhaling air that’s contaminated with droppings, urine, or saliva from infected rodents. But the particular strain identified in the cruise ship cases, called the Andes virus, can spread between people.
Health officials in several countries are working to trace the contacts of 29 people who disembarked the ship on the remote South Atlantic island of St. Helena on April 24, about two weeks after the first hantavirus death occurred. A Swiss man who left the ship early has tested positive for the virus and is being treated, and two people in the UK are reportedly self-isolating after returning home. Six people from the US were among those who got off the ship.
“The Administration is closely monitoring the situation with U.S. travelers onboard the M/V Hondius cruise ship with confirmed hantavirus,” the Centers for Disease Control and Prevention said in a statement on Wednesday.
Yet experts say there’s no need to panic at this point.
“It doesn’t spread terribly well, so I don’t have any concerns of this being the next Covid,” says Steven Bradfute, an immunologist and associate director of the Center for Global Health at the University of New Mexico. “Most of the spread in the past with this virus has been with close contacts—people sharing a bed, people sharing food, that sort of thing.”
The virus doesn’t spread easily with casual contact, and asymptomatic spread—a major driver of Covid cases during the pandemic—is also less likely. The available data on the Andes virus suggests it’s most likely to be transmitted when somebody is visibly sick, Bradfute says. Symptoms include fever, muscle aches, fatigue, and dizziness, which can progress to coughing, shortness of breath, and difficulty breathing.
“That is actually really helpful, because it makes it a lot easier to do contact tracing and to identify high-risk individuals,” he says, though he cautions that outbreaks of Andes virus are uncommon, and just because the virus has behaved one way in the past does not mean it always will. “The infections have been rare enough that we can’t say that with certainty.”
One of those outbreaks occurred from late 2018 into early 2019 in Patagonian Argentina, stemming from a birthday party attended by around 100 people. Three people were the main drivers of the outbreak, which resulted in 34 cases and 11 deaths. The authors of a study who traced the outbreak in detail found that 26 of the 34 cases became sick after close contact with someone who was infected, including people who hadn’t attended the party. Six people were likely exposed to the virus via droplets or aerosols.
Tech
ChatGPT Has ‘Goblin’ Mania in the US. In China It Will ‘Catch You Steadily’
Are you even online in 2026 if you haven’t experienced the verbal tics of ChatGPT? It loves goblins, em dashes, and “it’s not A; it’s B” sentence constructions. But what you might not know is that the chatbot also has plenty of strange phrases it loves to say in Chinese, and they are driving Chinese users crazy.
ChatGPT does a decent job answering questions in Chinese, which is why it’s widely used in China despite being blocked by the government. But when users make a request, be it a math problem or an image-generation prompt, the chatbot loves to answer: 我会稳稳地接住你, which literally translates to “I will catch you steadily [when you fall].”
Catch … what? A more generous translation could be, “I’ll hold you steadily through whatever comes.” But to any native Chinese speaker, the expression is annoyingly affectionate and out of place. Sometimes, the model gets more effusive and says in Chinese: “I’m right here: not hiding, not withdrawing, not deflecting, not running. I’ll be steady enough to catch you.” Yes, the sound you just heard was millions of Chinese ChatGPT users rolling their eyes at the same time.
Today, this sentence is the most prominent example of many verbal tics that OpenAI’s models have exhibited when talking to people in Chinese. Another tic widely talked about on social media is how the model loves to say 砍一刀 (“Help me cut it once”), a maddeningly ubiquitous marketing slogan by PDD, a major Chinese ecommerce platform that also owns Temu.
The phenomenon where models latch onto a specific phrase and overuse them to the point that they feel forced is called “mode collapse,” says Max Spero, cofounder and CEO of Pangram, an AI writing detection tool. It’s usually caused by post-training where AI labs give LLMs feedback on their responses. “We don’t know how to say: ‘This is good writing, but if we do this good writing thing 10 times, then it’s no longer good writing,’” Spero says.
Becoming a Meme
The phrase “I will catch you steadily” comes up so often in ChatGPT’s responses that it has become a meme on the Chinese internet. One image depicts the chatbot as an inflatable rescue airbag, eagerly waiting to catch people as they fall.
Zeng Fanyu, a 20-year-old developer from Chongqing, China, tells WIRED the meme inspired him to develop an April Fools’ project called Jiezhu, or “catch” in Chinese. Jiezhu is an open-source-prompt engineering tool that helps chatbots understand a user’s intention. “The idea for Jiezhu was so funny that I had a lot of motivation when I was developing it,” Zeng says. When he used ChatGPT to help with coding, the chatbot once again used the phrase jiezhu in its responses, completely unprompted.
OpenAI is aware of the meme. When releasing its new image model in April, one of the sample images shared by the company actually made fun of the phenomenon. In the picture, which resembles a comic book, Boyuan Chen, a Chinese researcher at OpenAI, depicts himself looking frustrated that the new image model has once again learned to say the same phrase. “This sentence has been memed as an unnatural but funny Chinese sentence GPT likes to use on Chinese internet,” his prompt reads.
OpenAI didn’t immediately respond to WIRED’s request for comment.
Is It a Bad Translation?
There are two likely explanations for why ChatGPT has become obsessed with the phrase “I will catch you steadily.” The first is that it could be the result of an awkward translation.
Several people I spoke with noted the phrase has a similar meaning to “I’ve got you,” which makes sense as a catch-all response in English. But while “I’ve got you” in English reads casual and concise; “I will catch you steadily” in Chinese sounds wordy and desperate. One user also looked through their chat history to show me that the model often says jiezhu, the Chinese word for “catch,” in places where it likely meant to say “understand,” pointing to a potential misunderstanding of what jiezhu means in specific contexts.
-
Tech1 week agoAlmost half of UK businesses hit by cyber attacks | Computer Weekly
-
Business1 week agoPSX plunges over 4,800 points | The Express Tribune
-
Business7 days agoApollo Sports Capital and Tom Dundon make landmark $225 million investment in pickleball
-
Uncategorized1 week ago
[CinePlex360] Some plugins have failed to updat
-
Politics4 days agoIran weighs US reply delivered via Pakistan as Trump signals opposition to deal terms
-
Fashion7 days agoCanada’s Lululemon appoints Esi Eggleston Bracey to board of directors
-
Tech7 days agoThis Indigenous Language Survived Russian Occupation. Can It Survive YouTube?
-
Entertainment1 week agoDavid Allan Coe, country singer who wrote “Take This Job and Shove It,” dies at age 86
