[gcheong]

I can only speak for myself but as far as the first two the problem is that given the pressure and time constraints your success is generally dependent upon whether or not you’ve churned through similar enough problems in preparation so it’s not as much a test of your coding ability as simply pattern matching ability, which is important on some level, but not nearly as useful as, say, being able to research a solution to a novel problem and having a general sense of where to start. But for some reason we’ve equated the ability to do leet code problems with the ability to do the latter and now I’ve got to spend a significant amount of time churning through a bunch of coding exercises to prove myself instead of proving myself by doing real projects and being able to speak about them. As for take-homes, my main problem with them, other than the risk that you’re doing unpaid work, is that you often just submit your solution to the void without any chance to discuss your solution and the trade-offs you made, improvements you would implement given more time, etc. so you can imagine it quickly becomes an onerous time drain that doesn’t seem like a fair trade of time on your part if every company asks for one.