Grabbing quick and easy answers from the Google and SO leaves me more time for the really interesting problems in my experience so far. And it's the really interesting problems that make writing bespoke C/GPU kernels worth the time involved in their creation so double-good, no?
To be fair, sometimes I try to solve problems for an hour or so before going to the Google or SO, and when I find I derived roughly the same approach, I consider it a done-deal that it's probably the right way to solve the problem at hand.
I would love to see a coding interview where the interviewee has access to Google/SO. It's a far more realistic portrayal of their on the job capabilities IMO. It would also force interviewers to come up with more interesting questions.
To be fair, sometimes I try to solve problems for an hour or so before going to the Google or SO, and when I find I derived roughly the same approach, I consider it a done-deal that it's probably the right way to solve the problem at hand.
I would love to see a coding interview where the interviewee has access to Google/SO. It's a far more realistic portrayal of their on the job capabilities IMO. It would also force interviewers to come up with more interesting questions.