> "...how about we give you a project that more closely reflects problems we solve here day-to-day that you can solve in the comfort of your home on your own hardware in whatever IDE makes you comfortable?"
"How about you, a highly trained professional, do 5-10+ hours of unpaid work for us, that will be judged against desperate candidates who invested 25+ hours into them and/or bought a solution online, so you will get nothing for hours of work, except a meaningless rejection notice by some overworked programmer who didn't even read much of your code since he had 50 candidates' solutions dumped on his plate?"
How about working to improve those take home projects since it seems to be more promising than the whiteboard interview which is doomed to suck. For example, my company has a take home project that involves hitting some APIs and doing some stuff with the JSON they return. It shouldnt take more than an hour and we give a time limit of 3 hours just so people don't spend days on it.
Who decided that whiteboard interviews are inferior to takehome assignments?
Personally, I've hired many great candidates, and got several great jobs, via whiteboard interviews. It's not perfect but it's far superior to takehomes, which are completely one-sided, very hard to get right (by the recruiting company), and often are boring.
Whiteboard interviews and (even better) pair-programming with an interviewer are reciprocal and allow me as a candidate to learn about the company rather than just do some work for them for free. Often the tasks are more interesting too.
In my opinion, it is the takehome tasks that are "doomed". Their premise is flawed: "we don't want to invest an hour of our engineers' time in talking you worthless candidate, so we'll instead have you spend several hours coding up a task that our engineers will never take the time to check properly, since their time is more valuable."
In my opinion, the future is platforms interviewing.io, where candidates are vetted and top candidates get the minimal investment of an hour of the company's engineer time to pair-program, which is far superior and more reliable way of screening.
"How about you, a highly trained professional, do 5-10+ hours of unpaid work for us, that will be judged against desperate candidates who invested 25+ hours into them and/or bought a solution online, so you will get nothing for hours of work, except a meaningless rejection notice by some overworked programmer who didn't even read much of your code since he had 50 candidates' solutions dumped on his plate?"