The Rise of Open Weights; And the fall of commercial AI?
The frontier labs have grown at remarkable speed, yet the gap between their models and the freely available open weights alternatives — predominantly emanating from China — has all but closed. This post traces that story from the founding of OpenAI in 2015, through the LLaMA leak and DeepSeek's breakthrough, to a current crop of models that match frontier performance at a fraction of the cost, and asks whether the commercial labs can capitalise on the technology before it commoditises beneath them.