All published AI news and analysis.

Ben Recht argues that uncertainty quantification in forecasting is too narrowly defined, primarily relying on error bars and prediction intervals. He suggests a need for more imaginative approaches to better prepare for forecast failures.

Rome OS is a compounding agent operating system designed for recursive agents. It positions itself as an open-source alternative to platforms like Grok Bot and Meta's Muse, built with TypeScript.

OpenWorkBuddy is an open-source, local-first AI office agent designed as an alternative to Claude Cowork and Codex. It can generate PPTX, DOCX, XLSX, and HTML files from a single request, supporting various AI models and CLI tools.

The Story Skills repository offers JavaScript-based agent skills designed for end-to-end story writing in markdown. Packaged as plugins for Codex and Claude Code, it aims to streamline creative writing workflows.

The Claude Code Skills Marketplace has released version v33.8.2, offering plugins and skills for ADR-driven development, DevOps automation, ClickHouse management, semantic versioning, and productivity workflows. This TypeScript project focuses on enhancing AI agent capabilities for developers.

The ailang project, written in Go, has released version v0.52.5. This project explores the concept of language as a substrate for AI reasoning, moving beyond its use for human expression.

SkillScout, a macOS application written in Swift, has released version 1.2.0. This tool helps users identify the skills loaded by their coding agents, determine which agents can utilize specific skills, and track skill usage frequency.

PiG (Pi in Go) has released version v0.4.1, offering a Go implementation of the Pi coding agent. This project aims for parity with the original TypeScript codebase, focusing on faithful translation rather than a rewrite.

Stim has released desktop-v0.1.3, offering fast, isolated local environments for React Native and Expo development. This tool is designed for coding agents, featuring shared build caches and parallel worktrees.

Agent Teams AI is a TypeScript project enabling multi-agent systems where agents communicate and review each other's work. It offers a Kanban board interface for high-level command and oversight, supporting over 300 models and 200 LLM providers.