विज्ञान

By Greg Bensinger SAN FRANCISCO (Reuters) -Just days after putting on its annual devices and services showcase, Amazon is losing a vice president who helps oversee the division. Rob Williams, vice president of device software and services and a member of the elite internal group known as the S-team advising CEO Andy Jassy, left his post and will step down entirely from Amazon at year's end, according to a memo sent internally Thursday and reviewed by Reuters. Amazon confirmed the memo after a query from Reuters without providing additional detail. "Rob has had a big impact on the software and experiences of nearly all of the products we’ve created and shipped," said Panos Panay, the senior vice president of devices and services, in the memo. He said Williams decided to "retire from Amazon," but Reuters could not learn his future plans. Panay also announced the consolidation of several teams in his unit, including moving the Alexa Smart Vehicle team into the broader Alexa group. He said the changes were effective Thursday including the elevation of Tapas Roy, previously a vice president overseeing Fire TV product and engineering, to Williams' former post.  Williams will remain an advisor to Panay and on the S-team through the end of 2025, according to the memo.  It is rare for S-team, or senior leadership team, members to depart, as the group is regarded as a badge of honor within the company with unique access to the CEO. Williams joined the 29-member S-team in late 2022 and worked at Amazon for 12 years. He attended Amazon’s devices event on Tuesday in New York where the firm showed off an array of new products like refreshed Echo voice assistant speakers, color Kindle e-readers and Fire TV sets with improved screen quality. In an email, he pointed to a LinkedIn posting where he wrote that he had been working on his exit all year and remained on to see the recent device launches. He did not disclose his plans beyond 2025. The money-losing devices and services unit has undergone multiple rounds of layoffs as it pares back on products. A multi-year project to update its Alexa voice assistant with embedded generative artificial intelligence has rolled out slowly and has no clear path to profitability and Amazon is overhauling its Fire tablets with Android operating software for the first time, Reuters reported in August. Earlier this week, Williams in a LinkedIn post touted the debut of a new operating system called Vega that will be in use in forthcoming Fire TV devices, including a streaming stick available later this month. Amazon’s proprietary Vega is meant to replace Google’s Android software and promises faster speeds at lower prices. “No one else has anything like it,” said Williams in the post. (Reporting by Greg Bensinger; Editing by Chizu Nomiyama)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

(Reuters) -AI chip firm Cerebras Systems on Friday filed to withdraw, effective immediately, its plans for an initial public offering in the United States. Last year, Cerebras filed for an initial public offering on the Nasdaq. The offering was delayed by a U.S. national security review of a $335 million investment by G42, an Abu Dhabi-based cloud computing and AI company. Cerebras on Tuesday said it raised $1.1 billion, bringing the company's valuation to $8.1 billion. (Reporting by Prakhar Srivastava in Bengaluru; Editing by Alan Barona)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

(Reuters) -AI chip firm Cerebras Systems on Friday filed to withdraw, effective immediately, its plans for an initial public offering in the United States. Last year, Cerebras filed for an initial public offering on the Nasdaq. The offering was delayed by a U.S. national security review of a $335 million investment by G42, an Abu Dhabi-based cloud computing and AI company. Cerebras on Tuesday said it raised $1.1 billion, bringing the company's valuation to $8.1 billion. (Reporting by Prakhar Srivastava in Bengaluru; Editing by Alan Barona)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

(Reuters) -AI chip firm Cerebras Systems on Friday filed to withdraw, effective immediately, its plans for an initial public offering in the United States. (Reporting by Prakhar Srivastava in Bengaluru; Editing by Alan Barona)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Debora Ely (Reuters) -A group of suspected scammers in Brazil using Instagram ads featuring deepfakes of supermodel Gisele Bundchen and other celebrities appears to have made millions of dollars through online fraud, according to police investigators. Authorities arrested four suspects tied to the scheme this week and froze assets across five states in an investigation that found over 20 million reais ($3.9 million) in suspicious funds identified by federal anti-money laundering organ COAF. The operation marks one of the first attempts in Brazil to counter the growing use of artificial intelligence tools to manipulate images and videos of celebrities for online scams. Brazil's Supreme Court ruled in June that social media platforms can be held liable for criminal ads posted by users if they fail to act swiftly to remove the content, even without a court order.  Instagram owner Meta said that its policies forbid "ads that deceptively use public figures to try to scam people," and that it removes such ads "when detected." The company added in a statement that it has "specialized systems to detect celeb-bait, invests heavily in trained review teams, shares tips on avoiding scams and offers tools to report potential violations." Bundchen did not respond to a request for comment. She previously warned followers about fake videos circulating on social media using her image to perpetrate scams. The probe announced this week by police investigators in the southern state of Rio Grande do Sul began in August 2024 after a victim reported being misled by an Instagram ad showing an altered video of Bundchen promoting a skincare product.  Another ad used the supermodel's likeness to promise a giveaway of suitcases, with buyers paying shipping fees for items that never arrived.  "We identified that the criminal group carried out a series of other scams, involving deepfakes of other celebrities and false betting platforms," Eibert Moreira Neto, head of Rio Grande do Sul's cybercrime unit, told Reuters.  Investigators are pursuing charges of money laundering and online fraud. They noted that most victims lost small amounts, usually under 100 reais ($19), and did not report the crimes.  "That created a perverse situation in which the criminals enjoyed a kind of 'statistical immunity.' They knew most people would not report them, so they operated at scale without fear," Isadora Galian, from the cybercrime unit, said in a statement.  (Reporting by Debora Ely; Additional reporting by Bernardo Barbosa; Editing by Brad Haynes and Aurora Ellis)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Debora Ely (Reuters) -A group of suspected scammers in Brazil using Instagram ads featuring deepfakes of supermodel Gisele Bundchen and other celebrities appears to have made millions of dollars through online fraud, according to police investigators. Authorities arrested four suspects tied to the scheme this week and froze assets across five states in an investigation that found over 20 million reais ($3.9 million) in suspicious funds identified by federal anti-money laundering organ COAF. The operation marks one of the first attempts in Brazil to counter the growing use of artificial intelligence tools to manipulate images and videos of celebrities for online scams. Brazil's Supreme Court ruled in June that social media platforms can be held liable for criminal ads posted by users if they fail to act swiftly to remove the content, even without a court order.  Instagram owner Meta said that its policies forbid "ads that deceptively use public figures to try to scam people," and that it removes such ads "when detected." The company added in a statement that it has "specialized systems to detect celeb-bait, invests heavily in trained review teams, shares tips on avoiding scams and offers tools to report potential violations." Bundchen did not respond to a request for comment. She previously warned followers about fake videos circulating on social media using her image to perpetrate scams. The probe announced this week by police investigators in the southern state of Rio Grande do Sul began in August 2024 after a victim reported being misled by an Instagram ad showing an altered video of Bundchen promoting a skincare product.  Another ad used the supermodel's likeness to promise a giveaway of suitcases, with buyers paying shipping fees for items that never arrived.  "We identified that the criminal group carried out a series of other scams, involving deepfakes of other celebrities and false betting platforms," Eibert Moreira Neto, head of Rio Grande do Sul's cybercrime unit, told Reuters.  Investigators are pursuing charges of money laundering and online fraud. They noted that most victims lost small amounts, usually under 100 reais ($19), and did not report the crimes.  "That created a perverse situation in which the criminals enjoyed a kind of 'statistical immunity.' They knew most people would not report them, so they operated at scale without fear," Isadora Galian, from the cybercrime unit, said in a statement.  (Reporting by Debora Ely; Additional reporting by Bernardo Barbosa; Editing by Brad Haynes and Aurora Ellis)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Debora Ely (Reuters) -A group of suspected scammers in Brazil using Instagram ads featuring deepfakes of supermodel Gisele Bundchen and other celebrities appears to have made millions of dollars through online fraud, according to police investigators. Authorities arrested four suspects tied to the scheme this week and froze assets across five states in an investigation that found over 20 million reais ($3.9 million) in suspicious funds identified by federal anti-money laundering organ COAF. The operation marks one of the first attempts in Brazil to counter the growing use of artificial intelligence tools to manipulate images and videos of celebrities for online scams. Brazil's Supreme Court ruled in June that social media platforms can be held liable for criminal ads posted by users if they fail to act swiftly to remove the content, even without a court order.  Instagram owner Meta said that its policies forbid "ads that deceptively use public figures to try to scam people," and that it removes such ads "when detected." The company added in a statement that it has "specialized systems to detect celeb-bait, invests heavily in trained review teams, shares tips on avoiding scams and offers tools to report potential violations." Bundchen did not respond to a request for comment. She previously warned followers about fake videos circulating on social media using her image to perpetrate scams. The probe announced this week by police investigators in the southern state of Rio Grande do Sul began in August 2024 after a victim reported being misled by an Instagram ad showing an altered video of Bundchen promoting a skincare product.  Another ad used the supermodel's likeness to promise a giveaway of suitcases, with buyers paying shipping fees for items that never arrived.  "We identified that the criminal group carried out a series of other scams, involving deepfakes of other celebrities and false betting platforms," Eibert Moreira Neto, head of Rio Grande do Sul's cybercrime unit, told Reuters.  Investigators are pursuing charges of money laundering and online fraud. They noted that most victims lost small amounts, usually under 100 reais ($19), and did not report the crimes.  "That created a perverse situation in which the criminals enjoyed a kind of 'statistical immunity.' They knew most people would not report them, so they operated at scale without fear," Isadora Galian, from the cybercrime unit, said in a statement.  (Reporting by Debora Ely; Additional reporting by Bernardo Barbosa; Editing by Brad Haynes and Aurora Ellis)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Kanishka Singh and Jaspreet Singh WASHINGTON (Reuters) -Apple said on Thursday that it had removed ICEBlock and other similar ICE-tracking apps from its App Store after it was contacted by President Donald Trump's administration, in a rare instance of apps being taken down due to a U.S. federal government demand. Alphabet's Google also removed similar apps on Thursday for policy violations, but the company said it was not approached by the Justice Department before taking the action. The app alerts users to Immigration and Customs Enforcement agents in their area, which the Justice Department says could increase the risk of assault on U.S. agents.  ICE has been a central part of Trump's hardline immigration agenda. Its agents have regularly raided and arrested migrants, and rights advocates say free speech and due process are often being infringed in the government's deportation drive. Apple's action may increase scrutiny of tech firms' growing ties to the Trump administration. Many companies, including the iPhone maker, have sought to avoid clashes with a White House that has not been shy about issuing threats — particularly around tariffs — against specific firms. "Based on information we've received from law enforcement about the safety risks associated with ICEBlock, we have removed it and similar apps from the App Store," Apple said in an emailed statement. Fox Business first reported the app's removal by Apple on Thursday. The Justice Department later confirmed that it had contacted Apple to pull the app and that the company complied. Google said its policies prohibit apps with a high risk of abuse. ICEBlock was never available on Google's Play Store. "ICEBlock is designed to put ICE agents at risk just for doing their jobs, and violence against law enforcement is an intolerable red line that cannot be crossed," U.S. Attorney General Pam Bondi said in a statement. Joshua Aaron, the Texas-based creator of ICEBlock, disputed that characterization and criticized Apple's decision. "I am incredibly disappointed by Apple's actions today. Capitulating to an authoritarian regime is never the right move," Aaron told Reuters. Now even launching a website would likely be met with takedowns, he said, adding that his legal team will decide on the next steps. Bondi has previously argued that Aaron is "not protected" under the Constitution and that they are looking at prosecuting him, warning him to "watch out." Civilian surveillance of federal immigration agents has grown more assertive since Trump returned to office, with activists saying they aim to protect their communities from aggressive ICE enforcement. In cities such as Washington, residents also rely on encrypted chats to share enforcement updates, though how often apps like ICEBlock actually alerted neighbors remains unclear. Six legal experts have told Reuters that surveillance of ICE is largely protected under the U.S. Constitution — as long as the activists don't interfere with that work. Courts have long held that recording law enforcement activities in public areas is legal. CHINA MAKES THE MOST DEMANDS Since Trump took office, ICE has raided multiple facilities with immigrants who are in the U.S. illegally, and has ramped up enforcement with $75 billion in new funding through 2029 for ICE. The agency has also arrested visa holders and permanent U.S. residents targeted by the Trump administration over pro-Palestinian advocacy. Legal experts have told Reuters that surveillance of ICE is largely protected under the U.S. Constitution, as long as they do not try to obstruct law enforcement. Apple removed more than 1,700 apps from its App Store in 2024 in response to government demands, but the vast majority — more than 1,300 — came from China, followed by Russia with 171 and South Korea with 79. Over the last three years, the United States has not appeared as one of the countries where apps were removed due to government demands, according to company application transparency reports. A majority of Apple's iPhones are manufactured in China, making the company particularly sensitive to tariff policies. The White House has floated potential taxes on imports of chips included in devices that would hit the company's substantial imports from China and other countries.  Apple removes thousands of apps from its App Store every year, including more than 82,500 in 2024, for other reasons, such as design-related issues, fraud or intellectual property infringement. (Reporting by Kanishka Singh in Washington and Jaspreet Singh in Bengaluru; additional reporting by Ted Hesson in Washington and Kenrick Cai in San Francisco; Editing by Edwina Gibbs, Anil D'Silva and Alan Barona)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Kanishka Singh WASHINGTON (Reuters) -Apple said on Thursday that it had removed ICEBlock, the most popular ICE-tracking app, and other similar software from its App Store after it was contacted by President Donald Trump's administration, in a rare instance of apps being taken down due to a U.S. federal government demand. The app alerts users to Immigration and Customs Enforcement agents in their area, which the U.S. Justice Department says could increase the risk of assault on U.S. agents.  ICE has been a central part of Trump's hardline immigration agenda. Its agents have regularly raided and arrested migrants, and rights advocates say free speech and due process are often being infringed in the government's deportation drive. Apple's action may increase scrutiny of tech firms' growing ties to the Trump administration. Many companies, including the iPhone maker, have sought to avoid clashes with a White House that has not been shy about issuing threats – particularly around tariffs – against specific firms. "Based on information we've received from law enforcement about the safety risks associated with ICEBlock, we have removed it and similar apps from the App Store," Apple said in an emailed statement. Fox Business first reported the app's removal on Thursday, citing a statement by U.S. Attorney General Pam Bondi, who said the Justice Department contacted Apple to pull the app on Thursday, and that the company complied.  "ICEBlock is designed to put ICE agents at risk just for doing their jobs, and violence against law enforcement is an intolerable red line that cannot be crossed," Bondi told Fox Business. Joshua Aaron, the Texas-based creator of ICEBlock, disputed that characterization and criticized Apple's decision. "I am incredibly disappointed by Apple's actions today. Capitulating to an authoritarian regime is never the right move," Aaron told Reuters in an emailed response. Bondi has previously argued that Aaron is "not protected" under the Constitution and that they are looking at prosecuting him, warning him to "watch out." Civilian surveillance of federal immigration agents has become more assertive since Trump returned to office, as activists say they are trying to protect their communities from aggressive enforcement by ICE agents. CHINA MAKES THE MOST DEMANDS Since Trump took office, ICE has raided multiple facilities with immigrants who are in the U.S. illegally, and has ramped up enforcement with $75 billion in new funding through 2029 for ICE. The agency has also arrested visa holders and permanent U.S. residents targeted by the Trump administration over pro-Palestinian advocacy. Legal experts have told Reuters that surveillance of ICE is largely protected under the U.S. Constitution, as long as they do not try to obstruct law enforcement. Apple removed more than 1,700 apps from its App Store in 2024 in response to government demands, but the vast majority – more than 1,300 – came from China, followed by Russia with 171 and South Korea with 79. Over the last three years, the United States has not appeared as one of the countries where apps were removed due to government demands, according to company application transparency reports. A majority of Apple's iPhones are manufactured in China, making the company particularly sensitive to tariff policies. The White House has floated potential taxes on imports of chips included in devices that would hit the company's substantial imports from China and other countries.  Apple removes thousands of apps from its app store every year, including more than 82,500 in 2024, for other reasons, including design-related issues, fraud or intellectual property infringement. Apple shares were down fractionally on Friday. (Reporting by Kanishka Singh in Washington and Jaspreet Singh in Bengaluru; Editing by Edwina Gibbs and Anil D'Silva)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)

By Mike Stone WASHINGTON (Reuters) -The much-needed modernization of the U.S. Army's battlefield communications network being undertaken by Anduril, Palantir and others is rife with "fundamental security" problems and vulnerabilities, and should be treated as a "very high risk," according to a recent internal Army memo. The two Silicon Valley companies, led by allies of U.S. President Donald Trump, have gained access to the Pentagon's lucrative flow of contracts on the promise of quickly providing less expensive and more sophisticated weapons than the Pentagon's longstanding arms providers.   Military drone and software maker Anduril boasted it had a prototype of the NGC2 communications platform working during a battlefield test just eight weeks after winning the contract award. But the September 5 memo provides fodder for critics who argue that Silicon Valley’s move-fast-break-things ethos may not be the best approach for vital military equipment.  The memo from the Army's chief technology officer about the NGC2 platform that connects soldiers, sensors, vehicles and commanders with real-time data paints a bleak security picture of the initial product. "We cannot control who sees what, we cannot see what users are doing, and we cannot verify that the software itself is secure," the memo says. Those concerns have been addressed already as part of the "normal process" of development, Anduril said. "The recent report reflects an outdated snapshot, not the current state of the program," the company said in a statement emailed to Reuters.  A Palantir spokesperson said "No vulnerabilities were found in the Palantir platform." However the Army internal memo written by Gabriele Chiulli, the Army chief technology officer authorizing official on the NGC2 prototype, said "given the current security posture of the platform and the hosted 3rd party applications the likelihood of an adversary gaining persistent undetectable access to the platform requires the system be treated as very high risk." Palantir stock was down more than 7% in afternoon trading on Friday. Anduril is not publicly traded, but company founder Palmer Luckey has said a public offering is planned. A September 30 article on Anduril’s website touted the NGC2 system’s performance during a live-fire exercise in Fort Carson, Colorado: "Soldiers fired 26 live missions with M777 howitzers on Fort Carson's live-fire ranges, running AXS side-by-side with legacy crews. The contrast was visible: one team struggling with delays, the other firing digitally in seconds." Despite the September 5 memo's critique, Leonel Garciga, Army chief information officer and Chiulli's supervisor, said in a statement to Reuters this week that the September 5 report was part of a process that helped in "triaging cybersecurity vulnerabilities" and mitigating them.  But the Army CTO's assessment, seen by Reuters and first reported by Breaking Defense, comes just months after Anduril was awarded a $100 million contract to create a prototype of NGC2 with partners including Palantir, Microsoft and several smaller contractors. The memo said the system allows any authorized user to access all applications and data regardless of their clearance level or operational need. As a result, "Any user can potentially access and misuse sensitive" classified information, the memo states, with no logging to track their actions. Other deficiencies highlighted in the memo include the hosting of third-party applications that have not undergone Army security assessments. One application revealed 25 high-severity code vulnerabilities. Three additional applications under review each contain over 200 vulnerabilities requiring assessment, according to the document. Palantir and Anduril are leading a new wave of defense firms aiming to transform the U.S. military with cutting-edge technologies, including drones, artificial intelligence, and autonomous systems. Palantir, a data analytics firm known for its work with intelligence agencies, and Anduril, which develops drones and battlefield software, have seen their valuations surge under Trump’s presidency, as Washington increasingly turns to Silicon Valley-style innovation to counter rising threats. Anduril recently signed a $159 million deal for an initial prototyping period to develop a night vision and mixed reality system as part of the Soldier Borne Mission Command (formerly IVAS Next) program. Palantir signed an $480 million contract for Maven, an artificial intelligence tool that sifts through images and sensor data to provide battlefield analysis. (Reporting by Mike Stone; editing by Chris Sanders and Sergio Non)

(The article has been published through a syndicated feed. Except for the headline, the content has been published verbatim. Liability lies with original publisher.)