In my experience, this is the norm in the ML scene. Giant globs of unreadable and in no way understandable code -- unless of course you already understand every

In my experience, this is the norm in the ML scene. Giant globs of unreadable and in no way understandable code -- unless of course you already understand everything. I think this is because the "product" so to speak is often the papers themselves, not the code, but I'm not sure.