Open source AI tools have reached quality levels that rival commercial offerings for many use cases. Here are the most important projects.
Large Language Models
Llama 3 (Meta) - The most capable openly available LLM. The 70B parameter model matches GPT-3.5 on most benchmarks. Runs locally via Ollama on a modern MacBook Pro or mid-range GPU.
Mistral AI - French startup producing high-quality open-weight models. Mixtral 8x7B is a mixture-of-experts model that punches above its parameter count.
Local AI Runtime
Ollama - The easiest way to run open source LLMs locally. One command installs and runs Llama 3, Mistral, or Phi-3. Perfect for developers who want AI without API costs or data privacy concerns.
Image Generation
Stable Diffusion (Stability AI) - Open source image generation that runs locally. The platform of community models, LoRAs, and ControlNet extensions makes it more customisable than any commercial tool.
ComfyUI - Node-based Stable Diffusion interface for power users. Supports complex multi-model workflows and automation.
Speech Recognition
Whisper (OpenAI) - Open source speech-to-text model that runs locally. Achieves near-human accuracy on clear audio. Available via local Whisper.cpp or hosted alternatives.
Vector Databases
Chroma and Qdrant - Open source vector databases for building RAG (Retrieval-Augmented Generation) applications. Essential for building AI applications on top of private knowledge bases.
