It is difficult to be sure from the flight data chart that was published in the Seattle Times, but it looks possible that the faulty AofA reading was below 25 degrees for most of the Lion Air flight.

Reading between the lines, Greenspun has presented these crashes as "the first mass killings by software". That claim looks clearer if you present it as wholly due to the absence of a few common-sense lines of code, and downplay the other systems design issues, such as the lack of triple redundancy, the failure to use the dual redundancy to detect sensor error, and the fact that the system was unnecessarily powerful, being able to drive the trim all the way forward.

From my pilot perspective, if a fellow pilot took the action MCAS did, I would consider it homicidal. This is based on the last instance of MCAS activation at ~05:43:21, airspeed is already exceeded maximum operating speed (Vmo), MCAS commands 1.3 units of nose down trim inside of 5 seconds, which changed the attitude to ~18 degrees nose down within 6 seconds. The vertical acceleration momentarily went negative, meaning everyone in the plane came off their seat, and anyone or anything not strapped down went to the ceiling. Soon after this we can see their maximum yoke nose up inputs up to that point, clearly this is the struggle point, and it was not effective, the mistrim was too great.

At the altitude and airspeed at the time this last nose down was commanded, it can lead to only death, it's not recoverable, so yeah mass killing by software is not an extreme claim.

If you view the MCAS as a control system that is a safety component of the 737 MAX, then you need a really good reason to completely cut off that safety system when a particular sensor input goes above a given threshold. In this case the argument being made is that if the input is above 24 degrees the safety system should be turned off, potentially leading to a stall. It's almost an argument against having that system entirely.

The problem is much more complicated than this and requires thinking end to end about 1) what the purpose of the system is, 2) when and how it should operating, 3) how much control it should have, 4) how it's activity is made visible to the pilot when it performs any control, 5) how and under what conditions it should automatically disable itself, or 6) be able to be manually disabled, and 7) how the pilot is made aware of all of these situations in a way that doesn't cause confusion in potentially complex scenarios involving other failures and alarms going off, and 8) proper training so the pilot can manage the plane when the characteristics have changed after the system is disabled. Oh, and 9) in this case since it's impossible for a pilot to manually trim a 737 when it's above a couple of hundred knots, making sure an electronic trim assistance function is available in this scenario (which it wasn't due to tragic idea to overload the use of the trim runaway cutout switches).

And I am probably missing a bunch of things, which is part of the point.

To be clear, I do not object to the 'verdict', but I do not think it helps to suggest that all the issues raised by these crashes can be dealt with just by giving each sensor one or two somewhat arbitrary thresholds beyond which they should be ignored. While I cannot think of any reason not to have adopted the proposed solution within MCAS, and it would have prevented the Ethiopian Airlines crash, it would not have eliminated the possibility in other circumstances (possibly including Lion Air), on account of the other problems with MCAS. If, in these cases, the sensors had failed to a plausible value (say, 15 degrees), the outcome would have been the same.