A Story of the Future of Open Source AI

AlexNet goes live in 2011 and is a game changer! AlexNet was the first neural net that showed a framework for GPU-based training and inference.

Google releases the now famous “Attention is All You Need” transformer paper in 2017. The transformer architecture is the modern underpinnings of generative AI.

OpenAI releases GPT3 in 2020 and GPT4 in 2023. Google and Anthropic follow quickly with Gemini and Claude, respectively. Meta launches an Llama, an open weight, but not open source model. None are open source, and only Llama’s weights are available.

We don’t know how any of these models were trained, or what systems were used to build them. No one small player can train models at these sizes from scratch. This is where Lumino comes in.

We’re building an open source AI research company so that everyone in the world can build and use AI in an easy, safe, and cheap manner. Hopefully, you have a sense of what we’re trying to accomplish here! Building what we’re building isn’t easy, so if this mission excites you, come join us!

Mission statement

Our mission is to enable the power of AI for every human. We do this by building infrastructure for AI - an open source, efficient, accessible, and cheaper infrastructure powered by blockchain and crypto.