An open-source AI investing framework called AI Berkshire, built on Claude Code and modeled after legendary investors like Warren Buffett, claims a verified 69% return in 2024 and 66% year-to-date in 2025, outpacing the S&P 500 by over 46 percentage points using multi-agent analysis and 16 specialized AI skills.
Orca launches a free, open-source AI Development Environment that lets developers run multiple coding agents like Claude Code and Codex simultaneously across desktop and mobile devices, all tracked in one unified interface under the MIT License.
A new GitHub repository called 'obsidian-skills' is taking the AI world by storm, teaching agents like Claude Code and Codex to interact with Obsidian vaults using Markdown, JSON Canvas, and CLI tools — and it has already amassed over 38,000 stars and 2,700 forks.
Baidu's newly released Unlimited-OCR, an open-source long-horizon document parsing model built on DeepSeek-OCR, is rapidly gaining traction with 5.6K GitHub stars, offering powerful single-image and multi-page PDF parsing capabilities with a live Hugging Face demo available now.
Cline launches a powerful open-source AI coding agent supporting every major AI model provider, offering developers full workflow automation, multi-agent coordination, CI/CD integration, and seamless connectivity across databases, APIs, and messaging platforms like Slack and Discord.
BuilderIO launches agent-native, an open-source framework enabling AI agents to operate directly inside real applications rather than as chat interfaces, offering shared actions, SQL-backed state, and a growing library of fully customizable SaaS app templates.
Stanford researchers unveil Proto, a groundbreaking AI-powered programming language that enables scientists to design complex biological systems — including DNA, RNA, and proteins — using natural language, achieving record experimental success rates and already validated in human cell lines.
Reflection AI, a young open-source startup founded by former Google DeepMind researchers, signs a massive $6.3 billion compute deal with SpaceX, securing access to cutting-edge Nvidia GB300 chips at $150 million per month starting July 2026.
A free, open-source GitHub project called LLM Algorithm Practice Lab is giving job seekers and AI practitioners hands-on Jupyter Notebook training in large language model development, covering everything from Python basics to advanced CUDA kernel programming through structured, test-driven exercises.
Stanford-spinout Radical Numerics emerges from stealth with $50M in seed funding and launches Omnii, a proprietary genome language model built to decode the 'grammar' of human DNA, as its earlier open-source models have already designed novel CRISPR systems and produced the first complete AI-designed bacteriophage genome.