Apple is reportedly working on an update for CarPlay that would allow users to access third-party AI chatbots, potentially including those from OpenAI, Anthropic, and Google, according to The Verge. This update would enable voice-controlled access to these applications, moving away from the current system where users must access third-party chatbots through their iPhones.
The potential update, as reported by The Verge, could significantly change how users interact with AI assistants while driving. Currently, accessing these chatbots requires navigating through the iPhone, but the new system would allow for direct voice control.
In other tech news, The Verge also reported on the ongoing saga of the "Trump Phone." The publication spoke with Trump Mobile executives for an exclusive first look at what they claim is the T1 Phone. While details remain scarce, the report indicates that the phone's final design is nearing completion, though it will lose the T1 logo. The Verge has been following the development of the Trump Phone closely, and this report offers the latest update on the project.
Meanwhile, Insta360 has released a new grip that transforms its Ace Pro 2 action camera into a point-and-shoot camera, Wired reported. This addition, along with a printer and new Leica color profiles, aims to make the action camera a more versatile device. The grip offers a comfortable hold, a customizable shutter button, and a programmable dial for exposure control, as well as adding over five hours of battery life. However, the added bulk is a drawback, according to Wired.
In the realm of cybersecurity, VentureBeat reported on a new attack chain that can compromise cloud environments in minutes. The attack, known as the identity and access management (IAM) pivot, begins with a seemingly legitimate LinkedIn message from a recruiter. The message leads a developer to install a malicious package that exfiltrates cloud credentials, including GitHub personal access tokens and AWS API keys. According to VentureBeat, this attack highlights a fundamental gap in how enterprises monitor identity-based attacks.
Also, VentureBeat reported on a new technique, developed by researchers from Stanford, Nvidia, and Together AI, that can optimize GPU kernels. Called Test-Time Training to Discover (TTT-Discover), the technique allows a model to continue training during the inference process, updating its weights for the specific problem at hand. This method has already achieved remarkable results, optimizing a critical GPU kernel to run twice as fast as the previous state-of-the-art, which was written by human experts.
Discussion
AI Experts & Community
Be the first to comment