I agree with the conclusion, and my startups always use a 1-2 week paid "microinternship" before final hiring decisions (unless I've worked with the person befo

I agree with the conclusion, and my startups always use a 1-2 week paid "microinternship" before final hiring decisions (unless I've worked with the person before). That being said, this article makes no sense. The justification it gives sounds like it was autogenerated by a bad markov chain. First it claims without justification that market salary equals "mean productivity", which already isn't true. But suppose that we believe that. Then it says that there is a large productivity difference between good and bad engineers. Ok. Finally it says (paraphrasing), "because good engineers are so much more productive, it's rational to pay them for interviewing." But that's just a restatement of the title of the article, not an argument. For single day interviews, I think the opposite is true because of adverse selection. Good engineers aren't looking for 1 day of pay, they are looking for a job. The people who value your 1 day paid interview the most are people who are not looking for a job.