Firstly, the author lost me when he stated, "Pilots today are learning their multiengine skills in airplanes like the Piper Seminole and the Beechcraft Duchess." When was that article written? Both of those airplanes entered production almost 45 years ago!
Prior to the introduction of the Seminole and Duchess (and continuing for decades afterwards) a lot of ME flight training was conducted in Piper Apaches, Aztecs, Seneca Is, and Twin Comanches, Cessna 310s, and Beech Barons and Travel Airs. Of those, the Apache, Aztec, and Seneca are relatively "benign," to use the author's term, whereas the Twin Comanche and Baron/Travel Air possess certain characteristics that proved to be problematic. Piper addressed the Twin Comanche's issue by increasing the published Vmc by, I think, about 10mph. I'm not aware of any such efforts to tame the Baron/Travel Air. The Baron has much in common with the Travel Air but has more power and a different vertical stabilizer. Both Baron and Travel Air are somewhat notorious and Vmc demonstrations in them should only be undertaken with great care. My understanding is that the Army had used B55 Barons as ME trainers and lost quite a few to Vmc-related loss-of-control. It may have been the Baron that the author had in mind when he wrote the article you refer to, and I agree based on my own experiences that the Baron/Travel Air can be less forgiving than the Seminole or Duchess.
That said, it's a stretch to claim that "...the critical nature of the VMC event has been downplayed in recent years." The point of the Vmc demonstration is not to exceed the limits of controllability--the point is to recognize the behavior of the aircraft as it approaches the limit of controllability and to take immediate corrective action before control is lost. Therefore the author's implied connection of "benign" and "downplayed" is tenuous, as the goal of the maneuver can be realized in any of the named aircraft. Practicing Vmc demonstrations in an airplane that exhibits "very violent VMC characteristics" seems rather unwise when other, more docile trainers are available--trainers that allow Vmc demonstrations to be conducted as intended while still being forgiving enough to allow recovery from typical student mistakes without viciously snapping into an inverted spin. When necessary to demonstrate Vmc in an airplane like the Baron, instructors will often employ techniques such as blocking the rudder pedals to limit rudder travel in order to artificially make the maneuver more benign. This practice effectively makes the Baron more like the Seminole and Duchess, which tends to contradict the author's point while at the same time increases the instructor's chances of actually cashing the paycheck that he risked his life to earn.
To answer the specific question you asked, every twin is different. I haven't flown them all, nor have I explored actual Vmc behaviors in many most that I have flown. I can generalize by stating that, within my limited sample set, higher-performance twins tend to react more abruptly and therefore are more demanding of timely corrective action. I have not experienced a stall of the vertical stabilizer--at least, not that I recognized as such--or a full-on departure from controlled flight during a Vmc training maneuver. It is worth remembering that the Seminole and Duchess were conceived and designed to be safe, forgiving trainers, whereas aircraft such as the Aerostar and MU-2 were conceived and designed to carry a certain payload a certain distance at a certain speed for a certain cost. The only Vmc demonstration I have personally, um, attempted, in a truly high-performance aircraft was done using a simulator (you can't get much more benign than a box that's bolted to the floor). I doubt that the flight characteristics of the simulator accurately mimic the actual aircraft in that flight condition, so no conclusions can be drawn.
