Open Source AI

The Allen Institute for AI (AI2) just dropped SERA, a new family of open-source coding agents that compete with massive closed systems. We break down their clever and surprisingly efficient training method that could change the game for repository-level AI.

Google just dropped TranslateGemma, a new family of open-source AI models specifically trained for high-quality translation. We break down how they work and why a smaller, specialized model can now outperform a much larger one.

A new open-source project called SETA is making waves by providing a complete environment for training and testing AI agents that live in the command line. Let's break down why this is a bigger deal than it sounds.

Ever wonder how AI could understand the world like we do? Meta's new PE-AV model is a huge step forward, learning to connect sights, sounds, and text in a single, unified brain. Let's break down how it works and why it's a big deal for the future of AI.

Ever wondered how an AI *really* thinks? OpenAI just released a new set of tools called 'circuit-sparsity' that lets us peek inside the black box, revealing the tiny, understandable circuits that power complex models. Let's break down what this means.

Ever wanted an AI creative partner that doesn't need an internet connection or expensive API keys? Let's walk through building a completely local storytelling pipeline using Griptape and a lightweight Hugging Face model.

Mistral just released Devstral 2, a pair of powerful AI coding models. One is small enough to run on your laptop, offline and private. But there's a major catch in the license you need to know about.

Chinese AI startup Zhipu AI just dropped GLM-4.6V, an open-source vision model that's more than just a chatbot that can see. It can actually use visual tools, a huge step forward for multimodal AI.

Cisco and Splunk just released a new open-source AI model for time series forecasting, and it's built differently. Find out why its unique "multiresolution" approach is a big deal for anyone working with real-world observability and security data.

For the past year, China has dominated the open-source AI scene. Now, a small US company called Arcee is fighting back with its new Trinity models, trained from scratch on American soil. Here's why this is a bigger deal than you think.

Chinese AI startup DeepSeek just released two incredibly powerful AI models for free. We break down how their new tech slashes costs and why it could seriously shake up the AI race between the US and China.

A new 1.5 billion parameter AI model from Weibo is making waves. Trained for just $7,800, VibeThinker-1.5B is outperforming models hundreds of times its size on complex reasoning tasks. Here's how they did it and why it matters.