> if they're too hard they can never be solved in 45 minutes, again yielding no useful information I strongly disagree with the assertion that you get no usefu

> if they're too hard they can never be solved in 45 minutes, again yielding no useful information I strongly disagree with the assertion that you get no useful information from an unsolved problem. The goal of giving a candidate a problem should not be to have them solve it. The goal should be to learn how they approach a problem, think it through, and work on solving it. Actually coming up with a solution is the least important part. As an interviewer I've intentionally given people problems that would take more time than they have. I'm always upfront about it, making sure they know they may not solve it and that they should get as far as possible while explaining their choices. As an interviewee I've been dinged for not solving problems. The reason why was because I spent most of my time talking through my approach and planning how I'd solve it. Without proper planning you get sloppy code, technical debt, and miss deadlines. Planning seems to be greatly undervalued in tech, especially in small startups.