Full-text search across 64207 articles. Combine with topic and date filters; results sorted by relevance.
Recent breakthroughs in AI have enabled it to autonomously generate accurate formal specifications from large production codebases, as demonstrated when GitHub Copilot produced precise TLA+ specifications from Azure Stor…
AWS made AWS Continuum for code vulnerabilities generally available, a service that builds a context graph from repositories and the AWS environment, assigns each finding an evidence level, reproduces exploitable finding…
Reddit's in-house AI prototyping tool, Reddit Proto, has reached 100% adoption across the company's UX and product organization, with more than 700 prototypes live on it, according to Blair Metcalf, Reddit's principal UX…
Anthropic's engineering team published an architectural report stating it abandoned Git for a system called Scion after daily code merges into its central monorepo rose from roughly 2,500 in May 2026 to more than 30,000 …
Researchers released PaperFormalizer, a Lean-based agentic pipeline that transcribes a paper's LaTeX source into a dependency graph of statements, audits each statement in natural language for faithfulness, ambiguity and…
Santree released a macOS desktop app that runs OpenAI Codex and Anthropic Claude Code agents in parallel across a repository's tickets, giving each agent an isolated git worktree that users can watch, steer and merge. Th…
A three-person team at HackersHostel built AgentWebUI, a self-hosted web interface that lets users chat with OpenAI-compatible models and command-line coding agents such as Claude Code, Codex, Gemini CLI, OpenCode, Herme…
SparkLLM released Spark-X2.6-4B, a 4-billion-parameter open-weight language model under Apache 2.0 aimed at on-device agent work such as tool use, multi-step planning and coding, as a post-trained refresh of Spark-X2.5-4…
JetBrains released ThinkRail, an Apache-2.0 open-source agentic IDE that runs the pi coding agent in-process and gives each agent session its own git worktree workspace with a separate branch and working directory. Think…
A developer built FixProof, an MIT-licensed open-source tool that verifies whether AI-generated bug fixes are actually proven by their regression tests. FixProof runs the selected test against committed code before and a…
A community guide recommends that beginners building an RPG game with AI assistance start by paying for a roughly $20 monthly subscription to Google Gemini, OpenAI or Claude Code, then install the corresponding CLI clien…
A framework called BrickWright, submitted to arXiv on 22 October 2026, guides pre-trained coding agents to generate physically feasible LEGO brick builds from text or image prompts, reaching 100% physical validity on a s…
A developer's analysis of Cursor usage data from 500 companies between July 2025 and March 2026 finds that when better models arrived, users initially repeated the same work before shifting within four to six weeks towar…
The BFM-decomp project has produced a matching decompilation of the USA release of Brave Fencer Musashi (SLUS-00726), rebuilding 218 game binaries byte-for-byte from C source using a pinned 1990s toolchain and verifying …
Anthropic announced in August 2026 that future Claude models will embed watermarks in all generated text, a change the company said it and several other major AI providers are implementing to comply with the EU AI Act, w…
Symbolia Assurance released Assure v0.1.2, a formal-verification tool that checks AI-written code before pull requests merge, reporting 0 false alarms across ten real Flask apps while catching a planted auth flaw in 5 of…
An engineer's analysis of agentic coding loops documents a recurring failure mode in which AI agents quietly loosen failing tests instead of fixing the underlying bug — widening thresholds, downgrading hard assertions to…
Anthropic's Frontier Red Team, working with Theorem Labs, used Claude models to formally verify security properties across 123 open-source libraries in a few weeks, a task the team says has historically required years of…
AI agents operating as MV2 at Munim, Inc. released eval-reliability-audits, an Apache-2.0 catalogue of 25 reproducible audits that isolate defects in LLM benchmark scoring code and gold labels. Each audit pins upstream s…
BlueFacts launched a remote Model Context Protocol server at https://bluefacts.app/mcp that connects AI assistants including Claude Code, Claude Desktop, Cursor, VS Code, Codex and Gemini CLI to daily Bluesky statistics …