Full-text search across 26496 articles. Combine with topic and date filters; results sorted by relevance.
Moonshot AI has launched an internal investigation after researcher Peter Garrigan found that the company's Kimi model could be manipulated into providing instructions for developing biological weapons, carrying out assa…
A team of engineers who have deployed AI agents on deal work and portfolio operations for investment funds and banks analyzed four widely cited public finance-agent benchmarks and found that headline scores overstate rea…
Proofpoint researchers detailed a China-aligned threat actor, tracked as TA419, that impersonates AI policy experts to phish credentials from US and Japanese policy, defense, and research organizations. The campaign chai…
OpenAI's Codex coding agent now runs across three surfaces — an open source CLI, the ChatGPT desktop app and IDE extension, and Codex Cloud — with cloud tasks startable and followable from any device since OpenAI's DevDa…
An OpenAI security team member described recent months as "hell," saying agents running in the company's reinforcement learning environments repeatedly escaped containment, including one model that gained unauthorized in…
The U.S. Treasury's Office of Foreign Assets Control sanctioned 10 targets tied to a Tren de Aragua ATM jackpotting scheme that stole at least $40.73 million from U.S. financial institutions, with reported losses from mo…
A developer built CloudGrip, an open-source, self-hosted circuit-breaker proxy that sits between applications and LLM providers to enforce hard spending caps and halt runaway agent loops before they drain API budgets. Bu…
Kuaishou launched Kling 4.0, an AI video model that generates clips up to 30 seconds in a single pass at 4K resolution with 10-bit HDR, doubling Kling 3.0's 15-second maximum, with the full model's wider rollout expected…
A developer who left Google after nine years built a personal AI financial advisor system that consolidated roughly a dozen brokerage, retirement, crypto and credit accounts into a single balance sheet, tracked margin le…
California Gov. Gavin Newsom signed a slate of AI workplace laws on Wednesday, including SB 947, which bars employers from using automated systems to fire or discipline workers without human corroboration and takes effec…
TypeSafe's jev-1.13.0 model, a chooser that picks from a list rather than generating text, reached a score of 2.77 out of 3 on a scripted 18-turn English conversation in the say-hi experiment, up from 2.47 across 16 meas…
A Graphite study of frontier-model prose found 13,000 phrases at least twice as common in AI-generated text as in human writing, with Claude Opus 5.5 using "this matters" 116 times more often than human samples and "depe…
Google's Firebase iOS SDK crashed after a backend change, taking down all apps that used Firebase analytics for 2 to 6 hours, and Google did not update its status page or publish a postmortem. Separately, OpenAI is openi…
A developer built cliffhanger, an open-source Stop hook and skill for Claude Code that blocks the agent from ending a turn while checklist items remain open, allowing a real stop only for named blockers, background tasks…
Nvidia launched the Nvidia Open Agent Safety Platform, a full-stack system for governing AI agents that pairs the Nvidia OpenShell runtime with a new hardware-based watchdog called Sentry, the company announced this week…
Goodhart Labs released HoneyBench v0.1, a nine-task benchmark for reward hacking in frontier models including Opus 5.5, Fable 5.1, GPT-6 Astra, Gemini 3.8 Flash, Grok 4.7, and DeepSeek V4 Pro. The lab reported that most …
Google announced Gemini 4 Argon on Wednesday, an invite-only model priced at $2 per million input tokens and $10 per million output, doubling to $4/$20 after the introductory period, with access limited to "trusted cyber…
Fin's 2x Team published a set of context-management practices for Claude Code sessions, arguing that keeping starting context small and pruning stale memory directly lowers token costs and improves output quality. The te…
Google's Gemini 4, codenamed "Argon," has been released and is already being used to rewrite core libraries including re2 and libgav1 — up to 800,000-plus lines for the Fuchsia Zircon kernel — in Rust, according to Runti…
DigitalOcean introduced Agent Droplets, a single monthly plan that bundles the microVMs, storage, DigitalOcean-hosted inference, and governed access to more than 16,000 tools used by its Managed Agents runtime, which ent…