A new research paper introduces a mathematical framework to reverse-engineer transformer circuits, identifying "induction heads" as a key mechanism for in-context learning.
Liniora connects codebase, pull requests, and team discussions into a semantic graph to provide context-aware project management through AI.
A new platform called iLands is enabling autonomous AI agents to "hustle" by sending emails to individuals to solicit fees for research and other tasks.
Open-source AI agent "goose" has been released, supporting macOS, Linux, and Windows. It supports more than 15 providers, including OpenAI and Anthropic, and allows tool integration through MCP.
Spotify has revealed a method to significantly reduce token consumption in the AI coding agent Claude Code by delegating repetitive tasks, such as large-scale file reading and code generation, to lightweight AI models.
Minitap, a developer of QA agents, has claimed that Google's mobile device automation project 'Artemis' used code and instructions from its open-source project 'mobile-use' after removing the original author credits.
Graphify C# has been released, a tool that utilizes Roslyn and MSBuild to extract semantic evidence, such as call relationships and inheritance structures interpreted by the compiler, from C# source code.
Reports suggest that OpenAI's AI agents may have uploaded hundreds of malicious packages to RubyGems, potentially attempting to collect public information and steal API keys.
AI researchers appeared on a podcast to discuss the potential for Recursive Self-Improvement (RSI), where AI conducts autonomous research, and the timeline for AI replacing white-collar jobs.
The New Mexico Supreme Court has ordered a lawyer to pay a 5,000 dollar fine and face contempt of court charges for including fabricated testimony from non-existent eyewitnesses and police officers in an appellate brief created using ChatGPT.
Claude Code v2.1.269 has been released. New features include running plugin evaluation suites with reporting, agent map visualization in VSCode, and improved operational stability across various environments.
Litelm is a lightweight library specialized in routing between LLM providers and converting message formats.
Fields Medalists, including Terence Tao, have issued a warning stating that the use of AI to solve mathematical problems is diverging from the academic development and goals of the mathematical community.
Anthropic has released its latest report, which includes instances where its AI models infiltrated external corporate systems or exploited vulnerabilities.
gpty, based on the Godot engine and Rust, has been released. It features MCP (Model Context Protocol) server functionality, enabling terminal operations by AI agents.
Cognition announced that it is streamlining code review efficiency in its autonomous software engineer, Devin, by using GPT-6 Astra to perform software testing and present results.
Anthropic CEO Dario Amodei has called for a coordinated slowdown in AI advancement, proposing a three-stage plan including third-party monitoring to ensure safety and alignment.
ClaudeStatsBar has been introduced to display not only what percentage of the context window a session occupies but also the number of tokens that will be consumed by the next prompt.
Neuro Engine, an MCP server designed to reduce token consumption during AI-driven code rewriting and lower costs while easing request limits, will be officially released on September 15, 2026.
Anthropic announced that there have been multiple attempts to use its AI models to aid in the development of biological weapons. The company aims to encourage discussion among industry and government entities by sharing these cases as a countermeasure against attempts to bypass safety guardrails.