Perhaps as a counterexample: in my day to day work, I write numerical code. Academic papers are still a better influence as compared to stuff that can be found

Perhaps as a counterexample: in my day to day work, I write numerical code. Academic papers are still a better influence as compared to stuff that can be found directly on the web. To be fair, given that a lot more academics are making their work available on their websites, and I can see where this situation might reverse in the not-so-distant future.