Microsoft warns of job‑themed repo lures targeting developers with multi‑stage backdoors

Microsoft says it has uncovered a coordinated campaign targeting software developers through malicious repositories posing as legitimate Next.js projects and technical assessments. The campaign employs carefully crafted lures to blend into routine workflows, such as cloning repositories, opening projects, and running builds, thereby allowing the malicious code to execute undetected. Telemetry collected during an incident…

Read More

Anthropic alleges large-scale distillation campaigns targeting Claude

Anthropic has accused three Chinese AI developers of running large-scale campaigns to illicitly extract capabilities from its Claude model to improve their own systems. The company claims DeepSeek, Moonshot, and MiniMax used a distillation technique, where a less capable model is trained on the outputs of a more advanced one. More than 16 million interactions…

Read More

Multi-token prediction technique triples LLM inference speed without auxiliary draft models

High inference latency and spiraling GPU costs have emerged as the primary bottlenecks for IT leaders deploying agentic AI systems. These workflows often generate thousands of tokens per query, creating a performance gap that current hardware struggles to bridge. Now, researchers from the University of Maryland, Lawrence Livermore National Labs, Columbia University, and TogetherAI say…

Read More

Snowflake extends Cortex Code CLI to dbt and Airflow to streamline data engineering workflows

Snowflake has extended support for Cortex Code CLI, its terminal-based AI coding agent, to dbt and Apache Airflow to help data practitioners streamline engineering workflows. “With this extended support, developers unlock secure, context-aware AI assistance within their preferred data engineering systems — empowering teams to work with data wherever it lives, and build, manage, and…

Read More