A self-propagating AI worm targeting Microsoft Copilot for Word was publicly disclosed today by Norwegian researcher Håkon Måløy, and every fix Microsoft shipped was bypassed within days. The attack h…
Microsoft shipped the Agent Framework Harness on July 22, a production-ready execution layer that turns an LLM into an autonomous agent with automatic tool-calling, planning, persistent memory, contex…
An open-source runtime called turbo-fieldfare, released by developer Andrey Mikhaylov, runs Google's Gemma 4 26B-A4B model in approximately 2 GB of RAM on any Apple Silicon Mac by streaming expert wei…
OpenTelemetry graduated as a CNCF project in May 2026, now sharing a tier with Kubernetes and Prometheus after 1.36 billion JavaScript API downloads and production deployments at Anthropic, Bloomberg,…
Anthropic released a beta feature for the Claude API on July 1, 2026, allowing developers to add or remove tools mid-conversation without invalidating the prompt cache, which can reduce input costs by…
On July 28, 1,178 employees from OpenAI, Anthropic, Google, Meta, and other frontier labs signed a letter asking the US government to help slow AI development, specifically targeting automated AI rese…
OfficeCLI, an open-source binary that gives AI agents deterministic control over Word, Excel, and PowerPoint files without Microsoft Office, has reached 22.9k GitHub stars. The tool, which runs on mac…
Google launched Gemini 3.5 Flash Cyber, a fine-tuned AI security model that found 55 vulnerabilities in the V8 JavaScript engine, generated a 100% reliable remote-code-execution exploit in a controlle…
Keycloak 26.7, released this month, adds production-ready OAuth 2.1 authorization for AI agents using the Model Context Protocol, enabling per-tool scopes and eliminating the default all-or-nothing ac…
JetBrains released Rider 2026.2 on July 22, introducing Agent Skills that expose IDE analysis capabilities—profiler data, coverage maps, refactoring intelligence, and framework knowledge—directly to e…
Data security company Cyera agreed to acquire non-human identity startup Oasis Security for approximately $1 billion, a deal that underscores the industry's recognition that AI agents are running in p…
Coursera announced a $100 million investment in LearnVector, an AI-native learning startup founded by its chairman and co-founder Andrew Ng, giving Coursera roughly one-third equity and a commercial p…
NVIDIA committed $5 billion to Safe Superintelligence Inc. (SSI), a startup with no products or published papers, in exchange for access to its research and a partnership on future compute platforms. …
OpenAI released GPT-5.6 on July 9 as three distinct models — Sol, Terra, and Luna — with a 1 million-token context window and different pricing, context recall behavior, and reliability profiles for p…
Over the weekend of July 26–27, Reddit users discovered that hundreds of shared Claude conversations and Artifacts, including medical records, clinical trial data, internal company documents, children…
PJM Interconnection, the grid operator for 65 million Americans across 13 states, will force data centers consuming 50 megawatts or more to switch to backup generators during grid shortages starting J…
Anthropic's Claude Mythos Preview discovered a practical key-recovery attack on the post-quantum HAWK-256 digital signature candidate in roughly 60 hours, reducing effective security from 264 to 238 o…
Dynamic Resource Allocation (DRA) is now generally available in Kubernetes, replacing the integer-based device plugin API with a declarative API that enables fine-grained GPU scheduling for AI workloa…
Perplexity's Personal Computer desktop agent launched on Windows on July 28, routing tasks across 20 frontier models and accessing local files alongside Office apps, competing directly with Microsoft …
Claude Sonnet 5, shipped June 30 by Anthropic, introduces a five-position effort dial that controls thinking tokens per request, with adaptive thinking enabled by default and a new tokenizer producing…