Full-text search across 26478 articles. Combine with topic and date filters; results sorted by relevance.
Princeton researchers Adithya Bhaskar, Jeffrey Cheng and Danqi Chen built Queen, a 4-billion-parameter language model that climbed from 1782 Elo to 2697 Elo over seven rounds of training, a gain of more than 900 points, …
A developer released Crowsnest, a private local macOS workspace that lets users run Claude Code and Codex sessions across multiple accounts, continuing work on a second subscription account when one hits its provider quo…
A developer tested how json-repair handles truncated LLM JSON output by cutting five sample JSON responses at every possible position, producing 853 truncated strings. While json.loads raised errors on all 853, json-repa…
TypeSafe's jev-browser-wingman MCP proxy completed all 34 browser-task runs in a 68-run comparison against Playwright alone, which completed 32 of 34 (94%), with both Playwright misses on the file-upload task. The proxy'…
Reflection AI Inc. introduced Beam, an open-source large language model with 501 billion parameters, which the company says matches or beats GLM-5.2 on some tasks while using between one third and one fourth the hardware…
The Wikimedia Foundation disclosed on October 5 that an internal investigation tied "rogue" OpenAI agent activity to a partial outage of its Wikidata Query Service that ran from May 7 to May 11, with more than half of re…
Google Mandiant's latest enterprise AI security report documents a single runaway agent that accumulated a $50,000 cloud bill, illustrating how agentic systems that loop on tool calls, retry failed API requests, or spawn…
Writing rules once in a repo-root AGENTS.md file lets Codex and Cursor read them directly, while Claude Code reads AGENTS.md only when no CLAUDE.md exists or via an @AGENTS.md import line, according to a blog post publis…
Postman rolled out its Passport credential-brokering system in 5 days to secure AI agent access to 42 production systems — 33 SaaS vendors and 9 internal — removing 278 API keys from circulation and leaving 0 credentials…
Cohere is offering a lockdown mode for AI agents that enforces strict access control lists, positioning the capability as an alternative to OpenClaw with built-in access control. The offering was detailed in an interview…
Reflection AI unveiled Beam on October 5, its first frontier open-weight model, which the Nvidia-backed startup says matches the reasoning performance of Z.ai's GLM-5.2 while being three to four times more efficient than…
Reflection AI, a New York-based startup founded in 2024 by former DeepMind researchers, released its first open-weight model, Beam, which CEO Misha Laskin said needs three to four times less computing power to reason thr…
Developer Gautam Khosla released mcp-pin 0.1.4, a local proxy that fingerprints every MCP tool's full metadata at approval time and blocks the session with a diff if any definition changes on a later connect, addressing …
Present-day AI systems are somewhere between median human and median professional human level at persuasion, according to a Forethought analysis published in September 2026, despite lab experiments showing frontier model…
A developer published a Next.js 15 streaming pattern for LLM chat apps using the Vercel AI SDK, showing that `streamText` with `toDataStreamResponse()` on the server and the `useChat` hook on the client handles token str…
A developer tested ten community safety and utility mods for Claude Code against 85 sessions, 882 prompts and 5,993 tool calls on version 2.1.288, finding that a guard hook which throws is silently skipped by default, so…
OpenAI launched GPT-6 Astra on September 3, 2026, and CEO Sam Altman apologized on X the next day after the rollout gave enterprise cybersecurity customers access ahead of Pro subscribers. OpenAI had shelved the prior GP…
A developer's tutorial series on building LLM applications explains how to add streaming to a chat feature so users see tokens as they are generated rather than waiting roughly ten seconds for a complete response. Using …
OpenAI will add an optional text watermark to ChatGPT outputs for European Union users and expose watermarking switches plus a Content Provenance API in its API project and organization settings, according to official Op…
ClickHouse announced that executable UDFs are generally available on ClickHouse Cloud across AWS, GCP, and Azure, four months after entering public beta. The GA release adds a Native runtime for precompiled, statically l…