Seed Diffusion explores how we’re pushing large language models toward a parallel future where AI learns faster, works smarter, and opens doors to entirely new possibilities

Inside our ML team’s week-long debate on OpenAI’s newly open-sourced GPT-OSS models versus Google’s Gemma3N architecture, from kernels and quantization tricks to efficiency, multimodality, and the quiet arrival of local AI’s future.

We’re thrilled to welcome Flat Capital as a new investor in Pieces. Learn more about this exciting partnership and what it means for the future of local-first AI.

We put the top AI tools for Python coding to the test, not just to see which writes code the fastest, but which actually feels good to use, fits into your workflow, and makes building in Python more enjoyable.

We’ve all heard it — “the future of AI is in the cloud.” But the real story is that the future might be smaller, closer, and more personal. From Small Language Models (SLMs) to local-first AI, agentic constellations, and even bio-inspired designs, the next big leap in AI isn’t about scale. It’s about building smarter, faster, more private tools that actually work for you.

Local AI isn’t a distant dream, it’s already here. From nano-models running offline to industry-wide shifts toward privacy-first architecture, discover why real breakthroughs are happening on-device, not in the cloud.