To clarify, since this seems to be a slight misreading of one of my comments.
The original one is
https://news.ycombinator.com/item?id=19568158
and was written at night, so there's some loopy phrasing in there I never corrected. Let me unpack some of it to fit the context here.
The pressure applied on the stick itself does not change. Rather, the function of how much deflection you get out of the plane per unit force applied to the stick changes.
Design regulations state an airframe for which this curve at any point goes negative cannot be certified as a civil transport airplane.
That means, there should be no point in the flight envelope where the sensitivity of the controls drastically changes. It should be a predictable increase in force required to get more deflection, all the way to stall.
With the MAX without MCAS, this rule is broken at high AoA when the engine nacelles start making lift.
MCAS is meant to compensate for this necessary breakage by inducing some mistrim.
The effect of mistrim in that case is to cause the controls to require more force to induce those last few degrees of deflection, using the AoA sensor as the primary indicator as to whether the system should activate, in order to bring the curve into compliance.
The problem, as hinted by the response to my post in that thread, is that an honest to god stick-pusher would be more akin to the system you'd want for that. Adding the AoA sensor as a single point of failure is insane, due to the consequences should that sensor start spewing garbage data in normal flight.
Most of those are realizations I came to after stumbling across a Royal Aeronautics Society interview of D.P. Davies, a British test pilot for the ARB, the aircraft certification authority for the U.K., which I believe I linked in the comment.
At the time I wrote it I was in the midst of a deep dive into the more arcane aspects of control systems. I have some literature tucked away somewhere that I was referencing, bit I'd need to dig rather deep in my browsing history to recover all the context.
An interesting historical note: MCAS is similar to a system McDonnell Douglas implemented in MD-11's (LSAS, Longitudinal Stability Augmentation System) for the same reasons as Boeing eventually implemented MCAS: to be able to claim that the new aircraft flew just like the old MD-10.
Makes one wonder if someone brushed the dust off something that really shouldn't have been emulated again after the merger.