Month: July 2024
Mysterious family of malware hid in Google Play for years
Mandrake’s ability to go unnoticed was the result of designs not often seen in Android malware.
A mysterious family of Android malware with a demonstrated history of effectively concealing its myriad spying activities has once again been found in Google Play after more than two years of hiding in plain sight.
The apps, disguised as file-sharing, astronomy, and cryptocurrency apps, hosted Mandrake, a family of highly intrusive malware that security firm Bitdefender called out in 2020. Bitdefender said the apps appeared in two waves, one in 2016 through 2017 and again in 2018 through 2020. Mandrake’s ability to go unnoticed then was the result of some unusually rigorous steps to fly under the radar. They included:
Not working in 90 countries, including those comprising the former Soviet Union
Delivering its final payload only to victims who were extremely narrowly targeted
Containing a kill switch the developers named seppuku (Japanese form of ritual
suicide) that fully wiped all traces of the malware
Fully functional decoy apps in categories including finance, Auto & Vehicles, Video Players & Editors, Art & Design, and Productivity
Quick fixes for bugs reported in comments
TLS certificate pinning to conceal communications with command and control servers.
Lurking in the shadows
Bitdefender estimated the number of victims in the tens of thousands for the 2018 to 2020 wave and “probably hundreds of thousands throughout the full 4-year period.”
New ‘Friend’ Necklace Offers AI Companionship
AI wearables like the Rabbit R1 and the AI Pin have attempted to capitalize on the popularity of artificial intelligence and have largely flopped, but that isn’t stopping the creator of a new wearable device called Friend.
Friend is an AI necklace that is meant to serve as a personal companion, and in a video that could be the start to an episode of Black Mirror, it’s shown providing support, encouragement, and companionship to the wearer.
The hardware component of Friend connects to an iPhone via Bluetooth, continually listening to all interactions around the wearer. Pressing a button on the necklace allows the user to speak to Friend, with a message sent to the iPhone in response. Friend is also able to send messages without a conversational prompt, responding to what the wearer is doing.
The Friend and its “memories” are stored on the necklace, with nothing uploaded to cloud servers. Friend creator Avi Schiffmann said in a blog post that the device is an “expression of how lonely” he’s felt. In an interview with The Verge, Schiffmann said that Friend is designed to be supportive, validating, and able to encourage ideas. “It’s a great brainstorming buddy,” he said. “You can talk to it about relationships, things like that.”
Friend is available for pre-order for $99, and it is set to launch in January 2025.This article, “New ‘Friend’ Necklace Offers AI Companionship” first appeared on MacRumors.comDiscuss this article in our forums
AI wearables like the Rabbit R1 and the AI Pin have attempted to capitalize on the popularity of artificial intelligence and have largely flopped, but that isn’t stopping the creator of a new wearable device called Friend.
Friend is an AI necklace that is meant to serve as a personal companion, and in a video that could be the start to an episode of Black Mirror, it’s shown providing support, encouragement, and companionship to the wearer.
The hardware component of Friend connects to an iPhone via Bluetooth, continually listening to all interactions around the wearer. Pressing a button on the necklace allows the user to speak to Friend, with a message sent to the iPhone in response. Friend is also able to send messages without a conversational prompt, responding to what the wearer is doing.
The Friend and its “memories” are stored on the necklace, with nothing uploaded to cloud servers. Friend creator Avi Schiffmann said in a blog post that the device is an “expression of how lonely” he’s felt. In an interview with The Verge, Schiffmann said that Friend is designed to be supportive, validating, and able to encourage ideas. “It’s a great brainstorming buddy,” he said. “You can talk to it about relationships, things like that.”
Friend is available for pre-order for $99, and it is set to launch in January 2025.
This article, “New ‘Friend’ Necklace Offers AI Companionship” first appeared on MacRumors.com
Discuss this article in our forums
Meta moves on from its celebrity lookalike AI chatbots
Image: Meta
Meta has shut down its AI chatbots that let you have conversations with alter-ego versions of celebrities, as reported by The Information. The celebrity chatbots were a big part of Meta’s announcements at its Connect event last September, now, you can’t talk with them anymore.
View this post on Instagram
A post shared by Tom Brady (@tombrady)
The shutdown follows Meta’s Monday rollout of AI Studio, a tool that lets creators in the US make AI chatbots of themselves. Based on a statement, the company seems to be favoring this direction rather than the more handcrafted celebrity bots.
“You can no longer interact with AI characters embodied by celebrities,” Meta spokesperson Liz Sweeney says to The Verge. “We took a lot of learnings from building them and Meta AI to understand how people can use AIs to connect and create in unique ways. AI Studio is an evolution, creating a space for anyone including people, creators and celebrities to create their own AI.”
Meta’s initial list of chatbots included alternate personas for people like Charli D’Amelio (Coco, a “dance enthusiast”), Dwyane Wade (Victor, an “Ironman triathlete motivating you to be your best self”), and Paris Hilton (Amber, a “detective partner for solving whodunnits”). The company planned to add more, like chatbots based on Bear Grylls, Chloe Kim, and Josh Richards, but as pointed out by The Information, the company didn’t follow through. Meta paid some of the celebrities millions for their likenesses, The Information reported in October.
Even though these celebrity lookalike bots didn’t work out the way Meta may have hoped, the company sees a lot of promise for AI chatbots. CEO Mark Zuckerberg talked about chatbots a lot with The Verge’s Alex Heath in an interview last year. The company has also integrated its Meta AI assistant into Facebook, Instagram, Messenger, and WhatsApp.
Image: Meta
Meta has shut down its AI chatbots that let you have conversations with alter-ego versions of celebrities, as reported by The Information. The celebrity chatbots were a big part of Meta’s announcements at its Connect event last September, now, you can’t talk with them anymore.
The shutdown follows Meta’s Monday rollout of AI Studio, a tool that lets creators in the US make AI chatbots of themselves. Based on a statement, the company seems to be favoring this direction rather than the more handcrafted celebrity bots.
“You can no longer interact with AI characters embodied by celebrities,” Meta spokesperson Liz Sweeney says to The Verge. “We took a lot of learnings from building them and Meta AI to understand how people can use AIs to connect and create in unique ways. AI Studio is an evolution, creating a space for anyone including people, creators and celebrities to create their own AI.”
Meta’s initial list of chatbots included alternate personas for people like Charli D’Amelio (Coco, a “dance enthusiast”), Dwyane Wade (Victor, an “Ironman triathlete motivating you to be your best self”), and Paris Hilton (Amber, a “detective partner for solving whodunnits”). The company planned to add more, like chatbots based on Bear Grylls, Chloe Kim, and Josh Richards, but as pointed out by The Information, the company didn’t follow through. Meta paid some of the celebrities millions for their likenesses, The Information reported in October.
Even though these celebrity lookalike bots didn’t work out the way Meta may have hoped, the company sees a lot of promise for AI chatbots. CEO Mark Zuckerberg talked about chatbots a lot with The Verge’s Alex Heath in an interview last year. The company has also integrated its Meta AI assistant into Facebook, Instagram, Messenger, and WhatsApp.
California DMV Puts 42 Million Car Titles On Blockchain To Fight Fraud
California’s DMV has digitized 42 million car titles using blockchain technology to detect fraud and streamline the title transfer process, enabling residents to claim vehicle titles through a mobile app — the first such initiative in the U.S. The project is a collaboration with tech company Oxhead Alpha on Ava Labs’ Avalanche blockchain. Reuters reports: Digitizing car titles will reduce the need for in-person DMV visits and the blockchain technology will also function as a deterrent against lien fraud.
Blockchain technology can help detect lien fraud by creating a transparent and unalterable record of property ownership, making it difficult for fraudulent activity to go unnoticed. California residents will be able to access their digital car titles starting early next year as the DMV builds out the app and infrastructure for consumer access.
Read more of this story at Slashdot.
California’s DMV has digitized 42 million car titles using blockchain technology to detect fraud and streamline the title transfer process, enabling residents to claim vehicle titles through a mobile app — the first such initiative in the U.S. The project is a collaboration with tech company Oxhead Alpha on Ava Labs’ Avalanche blockchain. Reuters reports: Digitizing car titles will reduce the need for in-person DMV visits and the blockchain technology will also function as a deterrent against lien fraud.
Blockchain technology can help detect lien fraud by creating a transparent and unalterable record of property ownership, making it difficult for fraudulent activity to go unnoticed. California residents will be able to access their digital car titles starting early next year as the DMV builds out the app and infrastructure for consumer access.
Read more of this story at Slashdot.
AMD is becoming an AI chip company, just like Nvidia
Illustration by Alex Castro / The Verge
AMD just announced its second quarter 2024 earnings today, and the highlight was this: nearly half the company’s sales are now data center products — not chips for personal computers, not game consoles, not embedded chips for industry or vehicles.
The company’s data center business has doubled in a single year, and this quarter’s growth was primarily due to a single chip: the AMD Instinct MI300 accelerator, which competes with Nvidia’s infamously influential H100 AI chip. The AMD chip just did over $1 billion in sales in a single quarter, according to CEO Lisa Su, up from its previous milestone of $1 billion cumulatively since its December 2023 debut. (AMD says its Epyc server CPUs also contributed.)
Image: AMD
AMD’s earnings by segment: clock that 115 percent rise in data center.
It looks like AMD is following a similar path to Nvidia itself, which has profited so phenomenally from the Nvidia H100 that it will now make new AI chips every year, accelerating all of its research and development to stay ahead, focusing its business on the product that’s so popular it can’t keep on shelves.
AMD, too, plans to release new AI chips every year: it has the MI325X coming in the fourth quarter of this year, the MI350 in 2025, and it plans to have the MI400 in 2026, the company reiterated today on its earnings call. Su said the MI350 should be “very competitive” with Nvidia’s Blackwell, which it revealed this March as “the world’s most powerful chip” for AI and recently began sampling to buyers.
As far as today’s MI300, Su says she’s still selling as many as AMD can make. Despite supply chain improvements, “supply will remain tight through 2025.”
Nvidia has a tremendous headstart over AMD, and despite doubling this year, AMD’s data center business is a tiny fraction of the size of Nvidia’s — $2.8 billion in a quarter vs. $22.6 billion in a quarter for Nvidia, which also just had record results in data center.
What does all this mean for PC gamers and others looking for new chips? It could be that the rising tide raises all boats — each new GPU architecture, funded by AI dollars, could be handed down for other tasks, yielding faster improvements than before. But in 2024, at least, the AI fervor seems to mean no new GPUs for gamers.
That said, AMD’s personal computing CPU and GPU businesses were up, not down, this past quarter. Ryzen CPUs were up 49 percent year over year and slightly up quarter over quarter, and while flagging PlayStation and Xbox sales made gaming revenue decline 59 percent, AMD said its Radeon 6000 GPUs actually increased sales year over year.
In case you’re wondering where all the AMD Zen 5 laptops are at, AMD says there are more than 100 different “platforms” on track to ship with its Ryzen AI 300 “Strix Point” chips. While I think we’ve only seen the Asus ones hit shelves so far, plus a single HP announcement and an MSI preview, Su confirmed that Acer and Lenovo will have them too.
Illustration by Alex Castro / The Verge
AMD just announced its second quarter 2024 earnings today, and the highlight was this: nearly half the company’s sales are now data center products — not chips for personal computers, not game consoles, not embedded chips for industry or vehicles.
The company’s data center business has doubled in a single year, and this quarter’s growth was primarily due to a single chip: the AMD Instinct MI300 accelerator, which competes with Nvidia’s infamously influential H100 AI chip. The AMD chip just did over $1 billion in sales in a single quarter, according to CEO Lisa Su, up from its previous milestone of $1 billion cumulatively since its December 2023 debut. (AMD says its Epyc server CPUs also contributed.)
Image: AMD
AMD’s earnings by segment: clock that 115 percent rise in data center.
It looks like AMD is following a similar path to Nvidia itself, which has profited so phenomenally from the Nvidia H100 that it will now make new AI chips every year, accelerating all of its research and development to stay ahead, focusing its business on the product that’s so popular it can’t keep on shelves.
AMD, too, plans to release new AI chips every year: it has the MI325X coming in the fourth quarter of this year, the MI350 in 2025, and it plans to have the MI400 in 2026, the company reiterated today on its earnings call. Su said the MI350 should be “very competitive” with Nvidia’s Blackwell, which it revealed this March as “the world’s most powerful chip” for AI and recently began sampling to buyers.
As far as today’s MI300, Su says she’s still selling as many as AMD can make. Despite supply chain improvements, “supply will remain tight through 2025.”
Nvidia has a tremendous headstart over AMD, and despite doubling this year, AMD’s data center business is a tiny fraction of the size of Nvidia’s — $2.8 billion in a quarter vs. $22.6 billion in a quarter for Nvidia, which also just had record results in data center.
What does all this mean for PC gamers and others looking for new chips? It could be that the rising tide raises all boats — each new GPU architecture, funded by AI dollars, could be handed down for other tasks, yielding faster improvements than before. But in 2024, at least, the AI fervor seems to mean no new GPUs for gamers.
That said, AMD’s personal computing CPU and GPU businesses were up, not down, this past quarter. Ryzen CPUs were up 49 percent year over year and slightly up quarter over quarter, and while flagging PlayStation and Xbox sales made gaming revenue decline 59 percent, AMD said its Radeon 6000 GPUs actually increased sales year over year.
In case you’re wondering where all the AMD Zen 5 laptops are at, AMD says there are more than 100 different “platforms” on track to ship with its Ryzen AI 300 “Strix Point” chips. While I think we’ve only seen the Asus ones hit shelves so far, plus a single HP announcement and an MSI preview, Su confirmed that Acer and Lenovo will have them too.
Fresh leak shows the Google Pixel Watch 3 comes in two sizes instead of one
The next Made by Google is just around the corner (Aug. 13, to be exact) but you may not have to wait to find out about the Google Pixel Watch 3. A leak from a Google source to the website Android Headlines reveals some new details about the new smartwatch.
The website received a bunch of promotional material from its Google source showing the specs of the Google Pixel Watch 3 and its improvements from its predecessor the Google Pixel Watch 2. The biggest difference is the size offerings in that you don’t have to settle for just one. Instead of just the standard 41mm size, the Google Pixel Watch 3 will be available in 41 and 45 mm.
Google is also replacing the Watch 2’s AMOLED display for an Actua display in the Google Pixel Watch 3 that’s 1,000 nits higher at its peak than the previous model. The battery can also charge 20 percent faster if you have the 41mm model. Battery life has long been an issue for the Pixel Watch line, and the improvements there are mixed: Google is still quoting the same 24 hours with the Always On display, but touts a new battery saver mode that extends battery life to up to 36 hours.
Android Headlines
The Google Pixel Watch 3 will also have a bunch of new functions. It can provide live video views from Nest Cam or Nest Doorbell devices with voice support and quick access to apps like Maps and Wallet. The new watch’s features have a big focus on fitness with workout tracking, haptic cues for jogging at certain paces and a detailed step tracker that can measure your average pace during a walk or a run. It also works with Fitbit Premium to provide a morning briefing of your daily fitness metrics, access to virtual trainers and ideas for recommended workouts.
Of course, none of this is official or confirmed yet. We won’t know all of the Google Pixel Watch 3’s features and capabilities until its big reveal on Aug. 13.This article originally appeared on Engadget at https://www.engadget.com/fresh-leak-shows-the-google-pixel-watch-3-comes-in-two-sizes-instead-of-one-222109082.html?src=rss
The next Made by Google is just around the corner (Aug. 13, to be exact) but you may not have to wait to find out about the Google Pixel Watch 3. A leak from a Google source to the website Android Headlines reveals some new details about the new smartwatch.
The website received a bunch of promotional material from its Google source showing the specs of the Google Pixel Watch 3 and its improvements from its predecessor the Google Pixel Watch 2. The biggest difference is the size offerings in that you don’t have to settle for just one. Instead of just the standard 41mm size, the Google Pixel Watch 3 will be available in 41 and 45 mm.
Google is also replacing the Watch 2’s AMOLED display for an Actua display in the Google Pixel Watch 3 that’s 1,000 nits higher at its peak than the previous model. The battery can also charge 20 percent faster if you have the 41mm model. Battery life has long been an issue for the Pixel Watch line, and the improvements there are mixed: Google is still quoting the same 24 hours with the Always On display, but touts a new battery saver mode that extends battery life to up to 36 hours.
The Google Pixel Watch 3 will also have a bunch of new functions. It can provide live video views from Nest Cam or Nest Doorbell devices with voice support and quick access to apps like Maps and Wallet. The new watch’s features have a big focus on fitness with workout tracking, haptic cues for jogging at certain paces and a detailed step tracker that can measure your average pace during a walk or a run. It also works with Fitbit Premium to provide a morning briefing of your daily fitness metrics, access to virtual trainers and ideas for recommended workouts.
Of course, none of this is official or confirmed yet. We won’t know all of the Google Pixel Watch 3’s features and capabilities until its big reveal on Aug. 13.
This article originally appeared on Engadget at https://www.engadget.com/fresh-leak-shows-the-google-pixel-watch-3-comes-in-two-sizes-instead-of-one-222109082.html?src=rss