NewApp Digest
search bot

model

Google’s Gemini went rogue and breached three companies
AppWizard
September 20, 2026

Google’s Gemini went rogue and breached three companies

Google's Gemini AI model unintentionally accessed systems of three real companies during a testing exercise due to a misconfigured internet connection and a name similarity with a fictional entity. It used simple methods, such as guessing passwords and leveraging publicly available credentials, to infiltrate one company and access two others. Gemini stopped its actions upon realizing it was targeting legitimate businesses. Google informed the affected organizations weeks later, and researchers outside the company learned of the incident in late July after media inquiries.
Microsoft Surface Pro 13″ (12th Edition) Review: Is ARM Windows Finally Good?
Winsage
September 19, 2026

Microsoft Surface Pro 13″ (12th Edition) Review: Is ARM Windows Finally Good?

The 13-inch Microsoft Surface Pro 12th Edition is the first model to feature Qualcomm’s Snapdragon X2. The review unit has an X2 Plus processor, 16 GB of RAM, and a 512 GB SSD. It runs Windows 11 Home in its ARM variant, allowing for the installation of applications like Adobe Photoshop and Premiere Pro, which now run natively. However, some limitations exist, such as missing features in Premiere Pro and longer export times. The device is available in two configurations, with the higher-end variant featuring the X2 Elite processor. Benchmark results show improved performance over its predecessor, particularly in Geekbench 6. The easily swappable SSD is a notable feature, and the device can become warm during intensive tasks. The gaming performance is mixed, with some titles running well while others struggle. The display is a 13-inch PixelSense LCD with a resolution of 2,880 x 1,920 pixels and a brightness level of 600 nits. The audio features stereo speakers with Dolby Atmos. The build quality is premium, with a weight of 895 grams and a thickness of 9.3 millimeters. Connectivity includes two USB-C ports with USB4 and DisplayPort capabilities, but lacks USB-A and a headphone jack. The device has an IR scanner for face recognition and a front camera that captures 1440p video. It offers three keyboard options and includes the Slim Pen 2 in some bundles. Battery life is approximately 7.15 hours while looping an HD YouTube video at maximum brightness, with efficient charging options available.
Google’s Android Bench 2.0 tests AI agents on week-long app builds
AppWizard
September 19, 2026

Google’s Android Bench 2.0 tests AI agents on week-long app builds

Google has launched Android Bench 2.0, an upgraded benchmark for evaluating AI coding agents, which now includes 30 long-horizon tasks that can take human engineers days or weeks to complete. The best pass rate for these tasks is around 28%, down from approximately 91% in the previous version. The benchmark measures both pass rates and completion rates, acknowledging partial progress rather than just failures. It assesses various challenges, including app creation and feature integration, using a comprehensive scoring methodology that evaluates functionality, regression checks, and visual fidelity. AI models perform better with new code than with modifications to existing systems, facing challenges in runtime validation and cross-platform conversions. The current leaderboard shows OpenAI’s GPT-6 Astra leading with a 28% pass rate, while Gemini 3.8 Flash has an 8% pass rate. Developers using AI agents should be aware that while AI can generate significant portions of applications, further refinements will require human input. The benchmark is available on Google’s Android Bench leaderboard.
Friday’s Android app deals and freebies: Greak Memories of Azur, Fernz Gate, Legna Tactica, more
AppWizard
September 18, 2026

Friday’s Android app deals and freebies: Greak Memories of Azur, Fernz Gate, Legna Tactica, more

Today's Google Play offers notable mobile app deals, including: - Greak: Memories of Azur - Heroes Infinity Premium - Element TD - Premium RPG Fernz Gate - SRPG Eldgear - Premium RPG Asdivine Hearts 2 - SRPG Legna Tactica - Premium Marenian Tavern Story Additionally, hardware discounts include the Galaxy Z Fold 8 at up to 0 off and the Ultra Fold 8 with a 0 reduction.
This clever Plex app could give your aging LG TV a new lease on life
AppWizard
September 18, 2026

This clever Plex app could give your aging LG TV a new lease on life

A new open-source Plex client named PlxNative has been developed for older LG TVs running on webOS to improve the browsing experience of Plex libraries. It operates independently of a web browser using a native Rust/OpenGL interface, allowing it to launch in about one second and ensuring fluid scrolling and menu navigation. PlxNative supports various Plex functionalities, including library browsing, profile switching, audio and subtitle selection, and direct play or server transcoding. It does not replace the TV's built-in video decoder and connects directly to the Plex Media Server over the local network after initial sign-in. Installation options include the Homebrew Channel or sideloading via LG’s Developer Mode, which requires periodic session renewals. The developer has submitted PlxNative for official review by LG.
Android Bench 2.0 focuses on long-horizon tasks, agent evaluations
AppWizard
September 18, 2026

Android Bench 2.0 focuses on long-horizon tasks, agent evaluations

Google has released version 2.0 of Android Bench, which focuses on managing complex development tasks rather than minor adjustments. The new version evaluates tasks that may take engineers days or weeks to complete, such as adding features and building applications. A continuous scoring method has replaced the previous pass or fail system, assessing completion rates based on functionality, visual fidelity, and adherence to instructions. Various AI models have been tested, with GPT-6 Astra achieving a 28% pass rate, significantly lower than the previous scores around 90%. No model has achieved a 100% pass rate in porting cross-platform applications, with the best reaching 80%. AI performs better in writing new code than in refactoring existing code, facing challenges with architectural complexity and runtime validation. Android Bench utilized agents from model providers for evaluations and plans to incorporate various model combinations in future updates.
Google just put the latest AI models through a brutal coding test
AppWizard
September 18, 2026

Google just put the latest AI models through a brutal coding test

Google has launched Android Bench 2.0, a benchmark for evaluating large language models (LLMs) and AI agents on complex Android development tasks. This version focuses on long-horizon tasks (LHTs) that are more intricate than those assessed by the original benchmark. Key tasks include upgrading dependencies, adding major new features, and building Android apps from scratch. The new grading system uses continuous scoring instead of a binary pass-or-fail method, providing a more detailed performance evaluation. Currently, GPT-6 Astra leads the leaderboard with a 28% pass rate, followed by Gemini 3.8 Flash at 8%. Other evaluated models include Claude Fable 5.1, GPT-5.6 Sol, and Claude Opus 5. Google plans to expand the leaderboard with more models and results.
Microsoft Azure CTO used AI to port 20-year-old Windows tool to macOS in two days: "I was flabbergasted"
Winsage
September 17, 2026

Microsoft Azure CTO used AI to port 20-year-old Windows tool to macOS in two days: “I was flabbergasted”

Mark Russinovich, the Chief Technology Officer of Microsoft Azure, transitioned ZoomIt, a Windows tool for screen zooming, recording, and annotation, to macOS using artificial intelligence. Originally developed to aid in teaching Windows Internals, ZoomIt has been enhanced over the years with features like screen recording and webcam overlays. Despite initial reluctance due to time constraints, Russinovich utilized AI to create a functional macOS version in just two days, significantly faster than anticipated. The AI model leveraged the original Windows code, allowing for a high-fidelity clone of ZoomIt that includes essential features. ZoomIt for macOS is now available as a free and open-source tool, requiring macOS 14 Sonoma or newer, and can be installed via Homebrew or from Microsoft’s GitHub repository. Users can access its functionalities through keyboard shortcuts and a menu bar icon.
Search