Open-Source AI Was Supposed to Eat the Rich; The Rich Are Still Here, Still Charging

The article examines why open-source AI models, despite rapid progress and community enthusiasm, have not displaced proprietary frontier models from OpenAI, Anthropic, and Google. It likely covers gaps in compute, data quality, alignment, and the sustained capital advantage of well-funded labs. The analysis suggests that while open-source narrows the gap incrementally, the big players retain structural edges in scale, safety, and productization.