[pnathan]

> having to prove they know things in an industry where a lot of people don't know things.

The number of people I have interviewed who have been unable to code fibonacci or fizzbuzz, even the number of people with "senior" in their resume, is genuinely remarkable. I can appreciate the tiresome nature of interviews, but someone having to demonstrate that are not simply fakers is, simply, critical to the hiring process.

For people who are "hired by network", that is a different story. But if someone is coming in through the front gate, I will insist on having some kind of technical scrutiny, in technical and documented form, rather than simple conversations.

Work sample tests, according to the literature HN member `tokenadult` has gathered, provide the best predictor of future work.

My druthers today is for senior engineers to provide two work samples: code and design. Neither _particularly_ long or _particularly_ challenging, in a time frame that is not onerous. If the engineer has a portfolio with recent work, that is, IMO, a suitable replacement for code. Again, the point is to elicit fakery and have a demonstration of being able to do good work.

[Viliam1234]

> The number of people I have interviewed who have been unable to code fibonacci or fizzbuzz, even the number of people with "senior" in their resume, is genuinely remarkable.

I believe you; I have met some. Luckily, neither fibonacci nor fizzbuzz requires 4 hours to code.