[miketwo345]

There are 3 main problems with the whiteboard tech interview. 1. It's not actually correlated to job performance. Google (and others) have done tons of studies on this, and the summary was basically, "Yeah, it's complete bs." 2. It is meant to show how a candidate would solve a problem, but it forces the candidate to follow a particular problem-solving path. Who in their right mind, when being first introduced to a new problem, cuts off all access to Google and other resources and tries to derive a solution from first principles?! Recent grads maybe, but not anyone who's been doing this for some time. It's not a good measure of how a candidate solves a problem if that's not the candidate's typical way of solving a problem. It's like to trying to see how good someone can ride a bike by handing them a unicycle. 3. We're the only industry that doesn't recognize success. At what point should the technical interviews cease? 5 years in? 10? Senior Engineer? Director? CTO? Can we really not tell if someone knows their stuff by just asking them about what's on their resume or by giving them a take-home project to submit?

The fundamental problem is trying to evaluate coding skill in an interview environment. The ability to regurgitate from Cracking the Coding Interview says nothing about whether a candidate can write well-documented, well-tested, maintainable code. It fails even more at determining whether a candidate knows when to write their own stuff vs when to use a library. It fails to determine how well a candidate can learn. All of those things are far more important skills to have.