AI News
Latest AI engineering news, updated daily.
Ai Engineering
GitHub Code Security Can Now Detect Bugs Beyond CodeQL
GitHub Code Security adds AI bug detection to extend code scanning coverage beyond CodeQL in ecosystems like Bash, Dockerfiles, Terraform, and PHP.
Github Code Security · Code Scanning · Codeql
Ai Engineering
OpenAI has Shut Down Sora and a Billion-Dollar Disney Deal
OpenAI is shutting down Sora, calling it a 'side quest.' The framing tells you where AI companies think the real value is.
Openai · Sora · Codex
Ai Agents
Kimi K2.5 Is the First Large Model on Cloudflare Workers AI
Cloudflare Workers AI now serves Kimi K2.5 with 256k context, tool calling, prompt caching metrics, session affinity, and batch inference.
Cloudflare Workers Ai · Kimi K2 5 · Agentic Workflows
Ai Agents
Cursor Cloud Agents Can Now Run in Your Own Infrastructure
Cursor self-hosted cloud agents are now GA, letting teams run agent execution in their own infrastructure while Cursor handles orchestration.
Cursor · Self Hosted Cloud Agents · Enterprise Security
Ai Engineering
Google's Lyria 3 Brings Song Generation to the Gemini API
Google added Lyria 3 to the Gemini API and AI Studio, letting developers generate songs with lyrics, structure controls, and image input.
Lyria 3 · Gemini Api · Google Ai Studio
Ai Engineering
Google Is Solving the LLM Memory Bottleneck with TurboQuant
Google Research published TurboQuant, a data-oblivious quantization algorithm that compresses LLM key-value caches to 3.5 bits per channel with zero accuracy loss and up to 8x speedup on H100 GPUs.
Google Research · Turboquant · Quantization
Ai Agents
Cloudflare Ships Dynamic Workers for AI Code Execution
Cloudflare shipped Dynamic Workers, an isolate-based sandbox that starts in milliseconds and uses a fraction of container memory, now in open beta.
Cloudflare · Dynamic Workers · Sandboxing
Ai Agents
ChatGPT Shopping Gets Visual Browsing and Product Comparisons
OpenAI rolled out richer shopping in ChatGPT with visual browsing, product comparisons, and an expanded commerce protocol for discovery.
Openai · Chatgpt · Product Discovery
Ai Engineering
Arm Launches First In-House AGI CPU
Arm unveiled its first production silicon, a 136-core data center CPU for agentic AI workloads, with Meta as lead partner.
Arm · Data Center Cpu · Agentic Ai
Ai Coding
Claude Code Gets Auto Mode for Uninterrupted Agent Runs
Anthropic launched Auto mode for Claude Code, a research-preview permissions feature that lets coding agents run longer tasks with fewer approvals.
Anthropic · Claude Code · Auto Mode
Ai Agents
Databricks Launches Lakewatch, Buys Two Startups
Databricks launched its Lakewatch AI security product in private preview and disclosed acquisitions of Antimatter and SiftD.ai.
Databricks · Lakewatch · Ai Security
Ai Engineering
LiteLLM PyPI Package Compromised by Supply Chain Attack
Malicious versions of LiteLLM on PyPI contained a three-stage credential stealer that harvested SSH keys, cloud tokens, and crypto wallets.
Supply Chain Security · Litellm · Pypi
Ai Agents
ServiceNow Ships a Benchmark for Testing Enterprise Voice Agents
ServiceNow AI released EVA, an open-source benchmark for evaluating voice agents on both task accuracy and spoken interaction quality.
Voice Agents · Benchmarking · Servicenow Ai
Ai Engineering
Gimlet Labs Raises $80M Series A for AI Inference
Gimlet Labs raised an $80 million Series A led by Menlo Ventures to scale its multi-silicon AI inference cloud.
Gimlet Labs · Ai Inference · Series A
Ai Agents
Anthropic Adds Desktop Control to Claude Apps
Anthropic launched a research preview that lets Claude use desktop apps in Cowork and Claude Code, with Dispatch task handoff from phone.
Anthropic · Claude · Computer Use
Ai Agents
WordPress.com Now Lets AI Agents Publish Posts
WordPress.com launched MCP write tools that let AI agents create, edit, and publish content on paid plans with user confirmation.
Wordpress Com · Mcp · Ai Agents
Ai Agents
Researchers Publish MCP-38 Security Taxonomy
Researchers released MCP-38, a 38-category threat taxonomy for Model Context Protocol systems as MCP security work expands.
Model Context Protocol · Mcp Security · Threat Taxonomy
Ai Engineering
IBM's Mellea 0.4.0 Adds Agent Tooling to Granite Models
IBM Granite announced Mellea 0.4.0 and three LoRA-based libraries for RAG, validation, and safety on granite-4.0-micro.
Ibm Granite · Mellea · Rag
Ai Agents
NVIDIA Ships Nemotron 3 Content Safety 4B for On-Device Filtering
NVIDIA released Nemotron 3 Content Safety 4B, a multilingual multimodal moderation model for text and images, on Hugging Face.
Nvidia · Nemotron · Content Safety
Ai Engineering
OpenAI Explains Codex Security’s SAST-Free Design
OpenAI detailed why Codex Security starts from repository context and validation, not traditional SAST reports, in its research preview rollout.
Openai · Codex Security · Application Security