3 ms·

> A resume will never tell you if a person is genuinely interested in uncovering unknown truths about nature. But a 30-minute conversation will. A 30-minute co

by l0b0 6y ago

> A resume will never tell you if a person is genuinely interested in uncovering unknown truths about nature. But a 30-minute conversation will. A 30-minute conversation with a friend might, but this is implicitly talking about an interview setting. How confident are you in your own ability to tell whether "a person is genuinely interested in uncovering unknown truths about nature" knowing that nobody has been able to come up with a similar formula for interviewing programmers, where there are at least in theory easily measurable knowledge levels?

TeMPOraL 6y ago
> knowing that nobody has been able to come up with a similar formula for interviewing programmers, where there are at least in theory easily measurable knowledge levels? There is a technique that's successful and widely used. It's somewhat unfashionable to talk about it in today's political climate, it requires doing some actual work, and it isn't a "formula" - but it works. The technique is exactly what TFA describes: talk to the candidate. Apply a domain-specific bozo filter (like FizzBuzz) if you must to narrow down the pool of candidates, but ultimately what you need to do is to get multiple people from the team for whom you're hiring to talk with the candidate for some hours, in a semi-structured way. Cue in accusations of bias. Sure, it's biased, but what's the alternative? Pretending you can codify worker competency and fit as a set of cheaply testable metrics that are independent of company/team specifics? Such process doesn't exist, and will never exist, because the problem is AI-complete. Instead, why not trust that the people on your team are decent human beings looking for their own interest? And in fact, a lot of companies do just that and happily chug along. But admitting to that doesn't make you an industry thought leader, nor can it be packaged into a service and sold.

by l0b0 6y ago

TeMPOraL 6y ago

> knowing that nobody has been able to come up with a similar formula for interviewing programmers, where there are at least in theory easily measurable knowledge levels? There is a technique that's successful and widely used. It's somewhat unfashionable to talk about it in today's political climate, it requires doing some actual work, and it isn't a "formula" - but it works. The technique is exactly what TFA describes: talk to the candidate. Apply a domain-specific bozo filter (like FizzBuzz) if you must to narrow down the pool of candidates, but ultimately what you need to do is to get multiple people from the team for whom you're hiring to talk with the candidate for some hours, in a semi-structured way. Cue in accusations of bias. Sure, it's biased, but what's the alternative? Pretending you can codify worker competency and fit as a set of cheaply testable metrics that are independent of company/team specifics? Such process doesn't exist, and will never exist, because the problem is AI-complete. Instead, why not trust that the people on your team are decent human beings looking for their own interest? And in fact, a lot of companies do just that and happily chug along. But admitting to that doesn't make you an industry thought leader, nor can it be packaged into a service and sold.