DeepMind could no doubt argue the same for openAI failing to play go and to not pursuing protein folding. Their several years behind on that one. :) Anyway, th

DeepMind could no doubt argue the same for openAI failing to play go and to not pursuing protein folding. Their several years behind on that one. :) Anyway, there have been several open source versions in the gpt3-3.5 range. Those are your base to build off of. Their's just little incentive to try to go participate in a moneysink competition with Google and Microsoft.