Full-text search across 26374 articles. Combine with topic and date filters; results sorted by relevance.
Security researchers have identified a prompt injection vulnerability in Google Gemini's voice assistant that allows attackers to embed malicious instructions within ordinary smartphone notifications. The assistant proce…
A developer built a persistent session system for the Critique Coding Agent API, eliminating the cold starts and context re-pasting that plagued earlier chained runs. The update keeps the E2B sandbox and OpenCode server …
Claude Code's /rewind command allows developers to roll back both conversation history and file changes to any earlier checkpoint in a session. The command reverts all file edits Claude made during the rolled-back portio…
Nvidia, Microsoft, Google and other major tech companies are introducing new chips, laptops and software designed to power AI agents that can autonomously complete complex, multi-step tasks on a user's behalf. The compan…
A team at Redwood Research developed a proof-of-concept pipeline that transforms benign Claude Code transcripts into synthetic sabotage trajectories for automated red-teaming of AI monitors. The approach allows AI compan…
A 2026 paper titled "The Price of Meaning" has formally proven that the entire class of AI memory systems—including retrieval-augmented generation pipelines, vector databases, and agent memory based on embedding proximit…
Axiom, a seven-month-old startup, solved all 12 problems on the 2025 Putnam exam, scoring 12/12 within the time limit and outperforming both top undergraduates and other AI systems. The feat demonstrates the company's "v…
Persistia, an open-source tool, now provides Claude Code with persistent memory by scanning entire projects—code, architecture, docs, and configs—on first run and storing the information in small, focused files with a na…
XAI, Elon Musk's artificial intelligence company, has paused hiring for specialist tutors intended to train its Grok chatbot. The hiring freeze follows the layoff of at least 500 generalist AI tutors and data annotators …
A new AI news aggregator called VNN launched on Hacker News, claiming to verify every source it aggregates in real time. The platform curates stories from global events, including a UC Berkeley study estimating 3,400 exc…
Microsoft unveiled seven in-house AI models at Build 2026, including its first reasoning model, MAI-Thinking-1, which benchmarks show is roughly on par with Deepseek V3.2. The company also introduced Frontier Tuning, a m…
Nvidia unveiled the RTX Spark superchip, a new PC processor developed with Microsoft, during a keynote at Taiwan's Computex conference on Monday. The chip, which features up to 20 Arm CPU cores and 128GB of unified memor…
Researchers from the University of Toronto demonstrated a prototype AI-powered worm that can autonomously spread across networks, exploit vulnerabilities across multiple platforms including Linux, Windows, and IoT device…
Microsoft announced seven new MAI models at its Build conference, including the flagship MAI-Thinking-1 reasoning model built with clean data lineage and no third-party distillation. The company released a 109-page techn…
On June 2, 2026, Hcompany released Holo3.1, the first production-grade computer use model family with quantized checkpoints—FP8, NVFP4, and Q4 GGUF—designed for fully local inference. The release enables GUI automation a…
On 31 May 2026, analysts are drawing parallels between the current AI boom and the late-1990s Dot-Com era, predicting an imminent "post-bubble" phase for artificial intelligence. The key question is which companies will …
An engineer discovered that stateless AI APIs like Claude, GPT, and Gemini force developers to resend entire conversation histories with every request, causing input token costs to balloon by 46x or more by the tenth mes…
Cisco Systems outlined its latest AI infrastructure strategy at Cisco Live 2026, focusing on supporting enterprise customers as they deploy agentic AI and frontier models. The company announced new networking, security, …
Enterprise AI agents have moved from demo environments into operational software in 2026, with vendors including Microsoft, Salesforce, IBM, ServiceNow, Google, AWS, UiPath, OpenAI, LangChain, and CrewAI competing across…
MSFM (MindStudio-Flavored Markdown) adds three annotation primitives — block annotations, inline annotations, and pointers — to standard Markdown, enabling the format to compile natural language specifications into full-…