Presumably you're aware that job interviews and the actual job are different things? At a real job no one cares if you use Stack Overflow because there's ample other data to assess your productivity (like...your actual productivity over the months and years that you're employed). The challenge with interviewing is trying to extrapolate performance from a much, MUCH shorter time horizon[1]. You're not testing the exact same environment, but you're using a different environment (the test/interview) to hopefully assess skills that are relevant in the environment you're optimizing for.
This is hardly a new concept: exams work pretty much the same way. There are other ways to assess candidates (like looking at their portfolios), but they have their own weaknesses, and testing skills in a constrained environment can still reveal things.
[1] And this is for the benefit of both parties: I've lost count of how many people I've heard complain about companies that have gauntlets of technical interviews.
Presumably you're aware that job interviews and the actual job are different things?
This is the core problem with tech hiring circa 2016. The sooner this idea dies, the better. It's surprising that more companies haven't noticed the most effective interviews are ones that resemble the company's day-to-day work as closely as possible.
This is hardly a new concept: exams work pretty much the same way. There are other ways to assess candidates (like looking at their portfolios), but they have their own weaknesses, and testing skills in a constrained environment can still reveal things.
[1] And this is for the benefit of both parties: I've lost count of how many people I've heard complain about companies that have gauntlets of technical interviews.