As someone who will be graduating in less than a year of PhD in comp sci., I would like to encourage you to not give up on all PhDs with such assumption. Like you, I am not impressed with the prospect of academia and research in general. But I honestly don't think a majority of CS phd students at decent (top 50) universities will fail fizzbuzz test as long as they have had undergraduate CS background. Some of us may fail to implement BFS/DFS/merge/quicksort during a 45 minute interview, but I think this is due to the interviewing process for tech jobs being somewhat broken (I saw my peers studying 3 or more months with interview prep books/sites and thought it's becoming ridiculous). A much better way to test a candidate would be to assign him/her a take-home assignment/project and let him/her work on it independently or along with your team, and judge his/her candidacy based on it (if your company can afford the extra time/resource for such process). I'm sure you'll notice those PhD folks from CS would not be as bad as you think they are.
In any case, I'd (for selfish reason) encourage you to be open-minded and give applicants with PhD background at least an equal chance. Personally, I never wanted a PhD, but I went for it because I needed to keep my immigrant status legal, and I couldn't pay for a master's degree out of pocket (if I enroll for a PhD program, I get a master's degree for free on the way). The downside of being a PhD student is that it is significantly more difficult for me to find my way back into software positions, which is something I wanted to do eventually; the HR usually throws away my resume with the assumption(s) that I am either only interested in research or am "overqualified" (or that I'd ask for more money), all of which are incorrect for a lot of PhD students, like me, stuck in academia for now. For all HR or recruiters reading this, I'd encourage you all to give PhD students a fair shot for the programming positions at your firms. You might find that these PhD students are more motivated or have more persistence/endurance (as most who survive the PhD experience usually have) or if you're lucky, are better prepared and thoughtful programmers.
Why is failing to implement BFS/DFS/quicksort in a 45 minute interview the fault of the interviewing process being broken?
The idea that someone has to study 3 months for an interview is ridiculous. I suppose for a PhD student I can understand as they probably need to refresh their memory, but if that's the case then you could argue they shouldn't be passing the interviews without refreshing their memory if we are talking about normal software development roles.
In any case, I'd (for selfish reason) encourage you to be open-minded and give applicants with PhD background at least an equal chance. Personally, I never wanted a PhD, but I went for it because I needed to keep my immigrant status legal, and I couldn't pay for a master's degree out of pocket (if I enroll for a PhD program, I get a master's degree for free on the way). The downside of being a PhD student is that it is significantly more difficult for me to find my way back into software positions, which is something I wanted to do eventually; the HR usually throws away my resume with the assumption(s) that I am either only interested in research or am "overqualified" (or that I'd ask for more money), all of which are incorrect for a lot of PhD students, like me, stuck in academia for now. For all HR or recruiters reading this, I'd encourage you all to give PhD students a fair shot for the programming positions at your firms. You might find that these PhD students are more motivated or have more persistence/endurance (as most who survive the PhD experience usually have) or if you're lucky, are better prepared and thoughtful programmers.