Open models will largely lose if they keep chasing closed frontier AI models; instead, open models should serve as complementary tools to closed agents
Markets, capabilities, cope, and bewilderment in the industrialization of language models. — Article voiceover