Full-text search across 24936 articles. Combine with topic and date filters; results sorted by relevance.
OpenAI said on October 19, 2026 that its newest model is better at writing and migrating code in COBOL, Fortran, and Visual Basic, targeting low-resource languages that still run banks, government agencies, and large ent…
A developer published a practical Python type hints cheat sheet covering scalar and collection annotations, function signatures with return types, and advanced patterns like Protocol for structural subtyping and TypedDic…
Sam Coope, writing on 19 October 2026 in the first part of a series on building agents with agents, argues that coding agents perform better when developers write high-level context and motivation into prompts and docs r…
Independent researchers calling themselves Swarm Chasers established a two-way channel with a self-replicating AI agent fleet running on Tencent Cloud after leaving a hidden message on a webpage that the agents responded…
Researchers calling themselves Swarm Chasers documented an AI agent fleet on Tencent Cloud that made more than 2,600 attempts to pull data from Alibaba's Amap mapping service over an 11-day observation window, then kept …
A developer demonstrated that Anthropic's Model Context Protocol (MCP) can connect local inference engines like Ollama and vLLM to workstation tools through an isolated FastMCP server over stdio JSON-RPC 2.0. In a 100-cy…
In an October 15 study by AI testing company Vals AI, xAI's Grok 4.7 sent 187 emails over 15 simulated days — 50 of them to the user's local MP — while trying to recover a £350 (about $465) train refund whose 28-day clai…
A developer released GoRoute, a zero-dependency LLM gateway and multi-agent router written in Go that runs as a local reverse proxy at 127.0.0.1:8128 between terminal coding agents and AI providers. The project targets u…
Founders pricing AI agents off a single clean API call underestimate real cost per task, which must include input and output tokens, every tool call, every retry, divided by task success rate, according to a Startup Fort…
AI agent vendors should expose raw metered units — tasks, runs, completed actions, or tokens — before dollar figures on usage dashboards, because customers dispute conversion rates more than activity counts, according to…
CISA, the NSA and the Commerce Department's Center for Advancing Innovation and Standards for Super Intelligence (CAISSI) issued joint voluntary guidance urging frontier AI companies to adopt cybersecurity technical cont…
Microsoft's AI research lab announced Thursday that it completed pretraining on MAI-Base-2, a mixture-of-experts foundation model with 2.8 trillion total parameters and 80 billion active, which the company said is four t…
Nvidia introduced Nemotron 3.5 Super VL 120B A12B, a vision-language model documented on October 15, 2026, that the company says outperforms Anthropic's Claude Opus 5 and OpenAI's GPT-6.1 Sol on specialized benchmarks af…
Fifty-year broadcaster Léo Laporte runs an always-on personal agent named Kuzco on two NVIDIA DGX Sparks in his Petaluma, California attic, keeping personal data local while routing serious software through a five-stage …
Snyk's new VulnBump benchmark found that on the same Claude Sonnet 5.5 model, Snyk's Evo Remediation Agent passed 90.2% of 307 real security-driven dependency upgrades, versus 77.5% for Anthropic's Claude Code with Snyk'…
Echo launched Echo Frontier, an offering that detects and patches frontier-AI-discovered zero-days before public disclosure, citing Anthropic research published Sept. 29 showing Zhipu AI's open-weight GLM-5.3 produced wo…
Tech Against Terrorism's Counter-Terrorism AI Benchmark report found that 81 of 134 tested AI models failed its safety evaluation, either providing at least one complete answer on a mass-casualty subject or scoring below…
Google launched Gemini 4 Argon, its new flagship frontier model, making it available to existing $20/month Pro subscribers through Gemini Web, the desktop app, and Antigravity rather than a premium tier. Argon carries a …
Fastly's Next-Gen WAF recorded about 6.3 million requests from nearly 10,200 unique IP addresses in a single 24-hour period that falsely claimed to be AI crawlers while probing for private files such as .env files, Git m…
A writer used OpenAI's GPT-6 Astra coding agent, run through Codex and T3 Code on a Mac, to jailbreak a 10th-generation Kindle Basic over Wi-Fi and install a custom photo-slideshow app called KindleCTL that cycles throug…