Underneath all this: although companies need talent, interviews are adversarial and not structured to create a supportive environment so candidates are at their best. Quite often the interview setup is to see how the candidate performs at their worst-- stressed, time pressured, on the spot, performing for fragments of time to a number of people while in what (feels|is) an asymmetrical power relationship.
> Interviews are adversarial and not structured to create a supportive environment so candidates are at their best.
This is a really broad stroke and it may even be true to a high degree. But I've had (and given) plenty of interviews where we engage a topic and just talk about it openly as if we were already colleagues. Maybe that's a Swedish thing?
I suspect the grueling interview process is the result of CS being overrun by people who cheated through their programming classes and/or exaggerated (lied) on their resume, things which have become extremely prevalent in recent years. Plus in the US you can't get a meaningful reference anymore because of fear of lawsuits (former employers will often only confirm employment). There's nothing left to trust. Worse, I have come across a number of CS graduates who can't write a working program to save their lives, though many (particular international students) seem to be very well-prepared to talk all day about the technologies and all the cool projects they supposedly did. Not that a surprise exam is a good solution.
The best interviews I've participated in (on either side of the table) have been when it's been more of a conversation, not an explicit evaluation. Yes, there's a power imbalance, and I'm not sure there's a way to fix that, but you can counteract it by treating interviews as a two-way street.
Interviewers don't need to be adversarial to determine if a candidate is going to be good for the role. Interviewees shouldn't be passive, and should be using the time to determine if the company is even a good fit for them.
