Agreed. I like the paper and I think it raises important considerations, but it does overstate its case a bit. I mean - people do learn new codebases after all. It's just more difficult to do than some think.