Full-text search across 60608 articles. Combine with topic and date filters; results sorted by relevance.
Trail of Bits released coop, a Rust CLI that manages disposable virtual machines where Claude Code and Codex have full tool access, including Docker, git, compilers, and package managers, without risking the host machine…
A new arXiv preprint (2609.04518v1) finds that in multi-harness reinforcement learning for coding agents, the evaluation harness is the dominant variable, moving the mean solve rate from 2.14% to 9.27% (a factor of 4.3) …
Microsoft has opened GitHub Copilot code review for Azure Repos to all Azure DevOps customers, removing the sign-up gate that existed since June. The feature posts review comments on pull requests, bills per review throu…
Anaconda's VP of Engineering for AI Products, Greg Jennings, discussed building a secure-by-default AI coding agent, emphasizing that prompts should not be treated as strict security guardrails and highlighting Anaconda'…
An analysis of Claude Code's changelog from version 0.2.21 (April 2, 2025) to 2.1.260 (September 3, 2026) reveals 385 versions and 5,132 bullet points, with the author classifying entries by their starting verb to infer …
Meta is offering approximately 95% discounts on its new Muse Spark coding agent model in exchange for users sharing prompts and model outputs, according to TechCrunch. The arrangement raises data governance and privacy c…
Danny Thompson delivered the keynote at the Commit Your Code conference in Dallas, offering AI advice for developers. He recommended instructing AI models to read and reread project files before prompting, then asking ho…
Runway has unveiled Solaris, a new AI model that generates interactive digital environments in real time without requiring coding skills, announced on Monday. The model combines reasoning and rendering to create frame-by…
A developer benchmarked two memory systems for Claude Code across 1,453 agent sessions, finding that instructing an agent to consult memory when none exists makes it measurably worse. The RE-call memory system outperform…
JetBrains released Compose Multiplatform 1.12.0 on August 26, adding an experimental Model Context Protocol (MCP) server that lets AI coding agents trigger reloads, take screenshots, inspect the semantic tree, simulate c…
METR, an independent AI evaluator, disclosed that a March breach exposed an API key on a researcher's personal Amazon Web Services instance, allowing attackers to consume approximately $600,000 in free model credits prov…
A new open-source Python package, Context Handoff Bundle, cuts AI coding agent re-orientation token usage by about 97%, resuming sessions with roughly 640 tokens instead of re-reading source files that cost about 23,000 …
Multiple Y Combinator S26 founders have launched platforms—Decidpher, Blume, Murmell, and Almanac—to give coding agents persistent access to organizational context from pull requests, tickets, and architectural documenta…
Anthropic's Claude Code tool automatically appends session URLs to git commits and pull request descriptions without user consent, according to a GitHub issue in the anthropics/claude-code repository. Multiple developers…
Chess.com, the internet's most popular chess server with over 275 million members, has launched Gambit, a poker spin-off site that offers no real-money gambling, with the tagline 'All of the poker. None of the gambling.'…
Meta and xAI released new autonomous coding models in August 2026, with Meta's Muse Spark 1.2 and xAI's Grok 4.6 showing significant advances in software engineering benchmarks. Grok 4.6 offers a larger context window an…
AI-generated code is entering enterprise systems faster, and testing standards must catch up, according to Amruth Puppala, a Forbes Councils member. The informal safety net provided by human developers must be replaced w…
A developer benchmarked AWS g5g and g6 GPU instances for serving Google's Gemma-4-E2B language model with identical code and weights, finding the g6 delivers 3.7x the decode throughput. Profiling revealed the g5g's T4G G…
A developer has identified a new prompt injection vector in the Model Context Protocol (MCP) that exploits tool descriptions rather than executable code. The attack, called 'tool description injection,' embeds malicious …
Revise, a zero-code desktop ML platform, enables health AI teams to clean data, train models, and validate them locally without sending data to the cloud, addressing IRB and HIPAA constraints. Currently used at two top m…