That's fair, but when your entire interview process is focused on algorithms and not real-world implementation, you're far more likely to hire the person who can implement quicksort but little of real value. And make no mistake: these engineers exist.

The truth of the matter is that most startups aren't Google or Facebook. A good percentage, perhaps the majority, have a CRUD application of some kind that, even if relatively complex or widely used, will realistically never require engineers to reinvent the wheel, let alone even understand the internals of the programming language the application is built on.

These startups should absolutely fear hiring an engineer who can't knock out a simple internal dashboard, integrate with a third party API, create a half-decent database schema or customize an instance of [insert popular open source CMS or ecommerce platform] more than they should fear hiring an engineer who can't implement his own quicksort.