[vmsp]

Some companies do this and pay the candidate for their time, regardless of outcome. I don't think there's much to comment there. Some don't pay the candidate. In that case, it's just a predatory practice to take advantage of the tough job market.

[SaucyWrong]

Just my personal take on this, but I’d happily perform real work for free instead of sitting for leetcodes and behavioral questions. I struggle with those formats a lot but have no problems shipping in a realistic problem domain.

That tradeoff makes enough economic sense to me personally, but to each their own.

I do agree that companies flush with profit should be able to offer a stipend though, and unwillingness to do so is a signal I use to evaluate them.

[vmsp]

Everyone prefers real problems. It's something you already know how to do instead of something you explicitly have to train for.

It doesn't change the fact that the real work could be an hour's exercise or longer remunerated work. This isn't an either/or scenario like you put it. Plus, for a fact, companies will happily have you doing both the leetcode and the take-home test.

[quietsegfault]

I would never perform work for free. What if my work has a hidden oopsie that ends up causing problems later? What if the code review is lax, and a bug ends up hurting someone? Unlikely scenarios, but things I’d be concerned about.

[setr]

I don’t understand how those concerns are alleviated having been paid

[quietsegfault]

If it’s paid, I understand the limits and liability.

[8note]

i cant imagine wanting to operate code written by said interviewees

it seems like such an obvious target for the likes of north korean saboteurs.

Interships already exist as "doing work as an interview"