Disagree with this on several levels.
One specific challenge comes as tech companies push more and more compensation to bonuses and other must-be-present-to-win approaches (RSUs, etc). If you leave shortly before a trigger date, you've essentially been working the entire previous year at a discount, since your contractual bonus for the time you worked will never be paid. If you continue significantly past a cliff, then your transition time is being worked at a discount.
If you give significant notice, one of two things can happen. If you give notice before the cliff, planning to depart after the cliff (that is, maximizing the percentage of the time worked where the company actually pays you what they agreed you were worth), the company can accelerate the departure schedule and avoid paying out; if you give notice after the cliff, you're inherently volunteering for discount work for a company you didn't even want to work for at full price!
In practice, I think at this point that companies that choose to put a large amount of compensation behind a cliff this way are responsible for understanding the consequences of that choice. If you pay 30% or my annual comp, and that of all my peers, on Monthuary 15th, then you should assume that you will get a cluster of resignations on Monthuary 16th each year, that those departures will happen in the standard two weeks, and that because they are clustered you will be unlikely to hand off as easily and fully as if they were scattered. But hey, you managed to screw some of my coworkers who had to leave mid-year for family reasons out of a few bucks!