The competition between OpenAI and Meta has intensified, with Meta successfully recruiting at least eight leading researchers from OpenAI during the past few weeks, as reported by WIRED. OpenAI executives, including Chief Research Officer Mark Chen and CEO Sam Altman, have reacted strongly to the situation. OpenAI is “recalibrating compensation” to retain top talent. OpenAI staff are working up to 80 hours weekly. OpenAI prioritizes AGI development over rivalry with Meta....

The new tools are now available, while more will be coming soon to Azure AI Studio for generative AI app developers, says Sarah Bird, Chief Product Officer of Responsible AI at Microsoft. Bird acknowledged that “Prompt injection attacks have emerged as a significant challenge, where malicious actors try to manipulate an AI system into doing something outside its intended purpose, such as producing harmful content or exfiltrating confidential data.” The announcement, which was made on March 28, added that these tools will be useful for business leaders who are “trying to strike the right balance between innovation and risk management” and “want to ensure that their AI systems are not generating errors or adding information” that can erode user trust....

Microsoft has been concerned about OpenAI prices and speed for its business users. Both companies have been building strong partnerships for years. Microsoft has invested over $13 billion in OpenAI, and many of its AI features use OpenAI models like GPT-4. Microsoft wants to integrate AI models from different companies to power its AI model copilot. The tech giant wants to reduce its dependence on OpenAI. Anonymous sources confirmed Microsoft wants to reduce costs and optimize its own AI models....

According to a press release, the two companies will establish a joint innovation lab staffed by experts from both organizations. TotalEnergies—which operates in over 120 countries—will contribute its extensive energy sector expertise, while Mistral AI will bring its advanced AI capabilities—the startup recently launched its first reasoning model, Magistral. The French giant TotalEnergies and the startup Mistral AI announced a new partnership to innovate and optimize energy consumption. The companies will focus on low-carbon solutions and the development of technologies to reduce CO₂ emissions....

First published: Mar 31, 2020 Updated 2 times since publishing With five kids ranging in age from 5 to 16, I had to find the best parental control app that was easy to manage and monitor. Using my kids as test subjects, I bought and used over 50 apps on their phones, tablets, and the family computer. Then I ranked them based on the features, strengths and weaknesses, and customer support....

Between June and December 2022, a group called Dark Pink launched numerous APTs. There were attacks against a number of Asian countries, including Vietnam, Cambodia, Indonesia, the Philippines, and Malaysia. Bosnia and Herzegovina, a country in Europe, was also attacked. “Group-IB’s early research into Dark Pink has revealed that these threat actors are leveraging a new set of tactics, techniques, and procedures rarely utilized by previously known APT groups. They leverage a custom toolkit, featuring TelePowerBot, KamiKakaBot, and Cucky and Ctealer information stealers (all names dubbed by Group-IB) with the aim of stealing confidential documentation held on the networks of government and military organizations,” said a Group-IB Malware Analyst....

The device is designed to offer a potential solution for individuals experiencing central vision impairment, a condition affecting millions globally. The technology uses unique lens arrays to project multiple images onto the retina. AI software adjusts images, improving vision clarity for individuals with macular degeneration. Clinical trials show a 50% improvement in reading ability for some patients. The smartglasses feature a unique optical system that projects multiple images to the retina using a specialized lens array....

The new offering is designed to provide federal, state, and local agencies with a secure and compliant way to access OpenAI’s advanced models, including GPT-4o, for a variety of applications. Agencies can deploy ChatGPT Gov in Microsoft Azure commercial or government cloud environments. The platform supports GPT-4o for tasks like coding, text analysis, and image interpretation. Over 3,500 government agencies have used OpenAI tools for tasks since early 2024....

Even as a writer who also deals with the artists-against-AI dilemma, I sometimes get distracted from the ethical debate and overlook the, frankly, aggressive ways some tech companies train their AI tools—using people’s copyrighted creations without permission and without providing easy ways to opt-out. But when I learned about the hundreds of thousands of artists migrating from Instagram to this new Cara app after Meta announced it’d be using public posts to train its new AI models, I woke up from my AI hypnosis....

The Worldcoin project aims to create a global digital ID system by encouraging people to have their irises scanned on its ‘Orb’ device in exchange for free cryptocurrency (Worldcoin’s own token, WLD). The Comissão Nacional de Proteção de Dados (CNPD) cited a high risk to its citizens’ fundamental right to personal data protection, necessitating this urgent intervention “to prevent serious or irreparable damage.” It raised concerns about “the impossibility of deleting data or revoking consent,’’ especially in the wake of over 300,000 Portuguese people sharing biometric data with Worldcoin....

Last Updated: Aug 3, 2025 ProXPN is a small and relatively unknown provider that’s been around since 2009, but despite its limitations, ProXPN passes as pretty recent provider within the average spectrum. Although ProXPN uses very powerful encryption keys (2048-bit encryption) to secure your VPN connection and supports OpenVPN protocol, however, they also use 512-bit key for the cipher which can be a liability. Nonetheless, you shouldn’t be worried about any kind of exposure while connected to the Internet, thanks to VPN Guard, a kill switch implemented within Windows and Mac OS ProXPN clients....

The Record reported that the breach impacted 422,424 people, compromising data such as Social Security numbers, addresses, phone numbers, and health insurance details. Payment card information and treatment records were not included. Over 400,000 individuals’ data was exposed in a ransomware attack on AAC. Sensitive data, including Social Security numbers and health insurance details, was compromised. Rhysida ransomware gang claimed responsibility for the attack. The healthcare provider, which operates addiction rehab facilities across eight states, started sending breach notification letters before the holiday season....

Last Updated: Aug 6, 2025 I found that while Red Shield has good security and privacy measures in place, it lags behind the competition in terms of speed, unblocking, and server options. The VPN only has 14 servers across 14 regions, but those servers are backed by AES-256 bit encryption, a comprehensive no-logs policy, and advanced security protocols. I experienced slow speeds and poor unblocking capabilities, but I was continuously impressed by the safety features packed into this small VPN....

According to a report by Cybernews, its team of experts suspects that some of these collections—which include over 30 datasets with an average of 550 million records—belong to cybercriminals. The datasets vary significantly in size and language, with some in Portuguese and Russian. Researchers determined that most of the data originates from various infostealers—malicious software used to harvest sensitive information. Cybersecurity researchers discovered an enormous breach exposing over 16 billion login credentials....

The gesture-recognition algorithm, developed by Veesion under the leadership of Benoît Koenig, uses in-store security camera footage to analyze movement patterns. The system generates brief video alerts to store staff when it detects suspicious gestures. Veesion uses AI to detect shoplifting through body movement, not appearance. System sends alerts to store staff when suspicious gestures are detected. One store halved inventory losses in three months using the tech. “The algorithm doesn’t care about what people look like....

In a recent internal memo Lütke made it clear that AI should be the first solution explored when tackling new tasks. AI usage is now required in all Shopify employees’ daily work. Performance reviews will include questions about effective AI use. Shopify has launched AI tools like Sidekick and Shopify Magic. — tobi lutke (@tobi) April 7, 2025 “Before asking for more headcount and resources, teams must demonstrate why they cannot get what they want done using AI,” he wrote....

Bluesky reached 10 million users and announced it this Sunday Millions of users migrated from X after Brazil suspended the app in the country The social media app now includes video features “If you’re reading this, you’re one of the first 10 million users on Bluesky!” states the message shared from the social media’s official account. The post included a message in English and Portuguese as millions of users from Brazil migrated to the platform after the Supreme Federal Court ordered the suspension of X....

Last Updated: Aug 7, 2025 Stream Via could not deliver on its promises of a secure, residential UK connection. I was unable to connect to my assigned IP address despite multiple days of troubleshooting with the support team. Even if it did work, this is a non-traditional VPN that lacks many security and privacy features that are essential for a modern VPN. Thankfully, there are plenty of options out there if you’re looking for a high-quality VPN that can provide a stable, reliable experience....

Chungin “Roy” Lee, along with cofounder Neel Shanmugam, launched the tool—now called Cluely—to assist job seekers during technical interviews. Columbia student suspended for building AI that helps cheat in interviews. Cluely AI gives real-time answers during interviews via invisible browser overlay. Tool now used for exams, sales calls; not officially for dating. The AI tool gives users live answers through an invisible browser overlay that reads the screen and listens to conversations in real time....

According to the paper, the scientists’ goal was to “predict which stars are most likely to host an Earth-like planet (ELP)” using AI. The team trained a Random Forest, a machine learning algorithm, to classify ELPs with a “not hosting” and “hosting” criterion. Swiss researchers built a machine learning model to detect Earth-like planets in planetary systems. The AI model found 44 planetary systems that host at least one planet with the required conditions....