Google's Gemma 4 finally made me care about running local LLMs

Google's Gemma 4 is a family of open-source, open-weight AI models in four sizes (E2B, E4B, 26B, 31B) designed to run efficiently on consumer hardware including phones and laptops. The key innovation is 'intelligence-per-parameter' — squeezing more capability out of fewer resources. You can run the smaller models on a phone via Google's AI Edge Gallery app (offline, no API keys needed) or on a laptop using Ollama. For everyday tasks like studying, code explanation, PDF Q&A, and brainstorming, the models perform well enough to be genuinely useful, with the added benefit of full data privacy since everything runs locally.