Earth Codes Data Commons - Phase One

Starting with data available at the Internet Archive, Earth Codes is embarking on a huge, ambitious, but doable moonshot: creating a large language model (LLM) trained on "regenerative data" -- data that many believe will allow AI to help us create a regenerative future for humanity and all species. This will cost millions of dollars. This is just the first phase where we are raising $250,000. This covers the initial Step 1: Data Curation · Step 2: Model Architecture · Step 3: Training at Scale · Step 4: Evaluation for specific data that may include biomimicry, scientific papers, indigenous knowledge (where specifically authorized), biophysics and more. The end goal is that AI tools will be created by many entities (e.g. NGO's, companies, governments, citizen scientists, indigenous communities, etc) which will help us visualize and build: regenerative cities, agriculture, industrial designs, fashion, food networks, architecture, transportation, etc.