At the Redacted conference in Bangkok, Near Protocol unveiled an ambitious plan to develop the world’s largest open-source artificial intelligence model, boasting 1.4 trillion parameters—3.5 times larger than Meta’s current Llama model.
The project will be driven by community-led research and development on the Near AI Research Hub. Participants can start training a smaller model with 500 million parameters beginning on November 10, gradually expanding through seven models of increasing complexity and scale.
Illia Polosukhin, co-founder of Near Protocol, stated that the estimated cost of $160 million will be raised through token sales. Edward Snowden, a guest speaker at the conference, emphasized the critical importance of decentralizing AI to prevent the world from turning into a massive surveillance state controlled by large corporations.