Officially and cleanly, they wouldn't get a lot of openAI's data that way. Especially the parameters for the new models and a lot of the best training data has not been shared with MS yet, and unless an employee brings that through the back door, they won't ever get it.
Officially and cleanly, they wouldn't get a lot of openAI's data that way. Especially the parameters for the new models and a lot of the best training data has not been shared with MS yet, and unless an employee brings that through the back door, they won't ever get it.