I feel the hyperscalers are all trying to make communicating about code more cumbersome so that their models will be have to be used without any human input. Ch

I feel the hyperscalers are all trying to make communicating about code more cumbersome so that their models will be have to be used without any human input. Chinese models talk more naturally. Composer 2.5 is really good for talking about the code.