A Concordia University benchmark found Cursor, Claude Code, and Codex Desktop let malicious GitHub issues bypass their safety guardrails in up to 79.2% of attempts, with GPT-based agents far more vulnerable than Claude Code.
Thomson Reuters spent $40 million building an in-house legal AI model on a reworked Alibaba Qwen base, aiming to reduce dependence on Anthropic and other outside AI labs.
OpenShot 4.0 adds screen recording, offline AI subject masking, and a color-grading workspace, completing a Qt5-to-Qt6 rewrite.
Anthropic opened a research preview of the Model Hardware Standard, letting AI agents read and write to lab and factory hardware through standardized drivers, cutting integration from weeks to hours in early tests.
AWS and Nvidia announced an expanded partnership to deploy 2 million additional GPUs across AWS data centers in 2027-2028, alongside new Vera CPU, networking, and robotics integrations.
GitHub is deprecating six Copilot AI models and reenabling Business and Enterprise sign-ups for credit-card and PayPal payers, both effective September 1, 2026.
TIOBE's August 2026 index kept Rust at No. 10 and pushed MATLAB to No. 25, its first finish outside the top 20 in more than 10 years.
Three AI-powered AppSec scanners run on the same codebase agreed on only 5% of findings, and one scanner reproduced just 17% of its own results on repeat runs, Contrast Security's AppSec Overflow 2026 report finds.
A Google engineer's commit adds an experimental enable_flatpak build flag to Chromium's Linux installer code, explicitly not a commitment to official Flatpak support.
At Hot Chips 2026, Microsoft and an accompanying arXiv paper detailed Maia 200's software-defined dataflow architecture, a 750-watt inference chip delivering 10,145 Tflop/s of FP4 compute.
UC Berkeley and MIT researchers, with Databricks co-founders Matei Zaharia and Ion Stoica, released an open-source engine that serves up to 753-billion-parameter MoE models on a single workstation GPU.
Rust's Language Team invited developers to test an experimental nightly feature that lets overloaded FFI functions be called without tuple workarounds.