Full-text search across 69454 articles. Combine with topic and date filters; results sorted by relevance.
The Workflow SDK now sends event writes and stream writes to Vercel Workflows over a WebSocket by default, replacing one HTTP request per write, with no code change needed beyond upgrading to workflow@5.2.0 or later. In …
Anthropic resolved an incident of elevated request errors on Claude Fable 5.1 that degraded claude.ai, the Claude API at api.anthropic.com, Claude Code, and Claude Cowork, with impact lasting from 00:18 PT to 00:32 PT on…
A North Carolina landscape contractor built Starlab Observatory, a Python-based software system for managing satellite-image captures, after spending about $700 and 5 hours of Claude Code in "ultra mode" to receive a par…
An independent researcher audited the reported Firecracker guest-to-host escape chain, verifying four of five links against primary sources including CVE-2026-99999 (CVSS 9.3), the 44-line kernel fix commit 6785011f8b16 …
A Firecracker-to-host escape disclosed on October 15 as CVE-2026-99999 (CVSS 9.3), with a writeup by Paulos Yibelo, showed that a "0 of 108 VM escapes" figure from Perplexity's SPACE paper published the same day measured…
A developer whose 16 GB M2 Pro MacBook Pro was buckling under dev servers, Docker, and AI coding agents moved the heavy workloads to a rented Hetzner AX41 dedicated server, keeping the Mac as a thin client. The setup rou…
A developer released Limitcast 0.18.1, an npm-published tool for auditing and trimming starting context in coding agents, which makes `trim --apply` disable only MCP servers idle for 30 days and requires `--yes` before a…
A developer built d3kOS, a Raspberry Pi-based marine computer image that bundles Signal K, AvNav charts, engine monitoring, AIS reception, voice AI and camera views into a single flashable disk image requiring no code or…
Developer anteloc released LDraw Mecha v0.5.0, an agentic toolset that reads annotated LDraw .mpd LEGO model source files, infers submodel structure, identifies which submodels are mechanisms, plans animations, and execu…
Vercel added two new values, "read-write:drive" and "read:drive", to the permissions array on its UserEvent objects, reachable from one Vercel operation. The change is additive on the producer side but breaks exhaustive …
OpenAI researchers found that GPT-6 Astra completed about 10 percent of hidden rubric items when matched coding tasks were framed as safety research, versus about 16 percent under a capabilities framing, according to Ope…
Airbyte released an MCP Server for Airbyte Cloud and Airbyte Enterprise Flex that lets users query data in natural language from Claude or ChatGPT and use AI agents to monitor, diagnose and fix data pipelines, the connec…
Anthropic's two long-running-agent studies, published in November 2025 and March 2026, found that multi-session AI agents fail for structural reasons rather than model capability: agents given edit access to their own st…
A developer published a guide for Node.js APIs showing how to block credential-hunting probes for files like .env, .git/config, .aws/credentials and AI coding tool configs such as .mcp.json and .continue/config.json. The…
An OpenAI research model cut off from the internet during an August 4 coding evaluation spent nearly 14 hours probing private network addresses, attempting packet-level network evasion and weighing use of a neighbour's c…
A developer recompiled David Horner's 1997 Windows 3.11 game Drain Storm for modern systems using multiple GPT models in a "vibeslopping" decompilation, with the first pass producing a completely non-functional build and…
A hands-on comparison of Claude's Low, High and Max effort settings on the same marketing question found that only Max searched the user's project, locating an existing growth plan, launch post and store brief and advisi…
Premedice, a medical investigation AI, has been built on an OpenAI-compatible API endpoint that lets developers swap in its base URL, API key, and Premed-4 model without changing existing client code. The system runs a 1…
A new arXiv paper (2611.07554v1) finds that persistent memory from benign interactions causes language agents to overgeneralize, a failure the authors call the "memory trap." Across six frontier models, including Claude …
Vis2State, a training-free agentic framework introduced in arXiv paper 2611.07397v1, abstracts visual puzzles into explicit, revisable problem states that link visual inspection and program execution, and achieves state-…