I have a less negative reaction than most commenters seem to. I am not phased by the snarky know-it-all attitude of the author, and it actually makes his point

I have a less negative reaction than most commenters seem to. I am not phased by the snarky know-it-all attitude of the author, and it actually makes his point more concise and clear IMO. I also can't help but appreciate the annecdote of abstraction hiding exactly what needs to be shown for performance computing considerations.