As background, I'm an analyst/number cruncher and quite rational.
I think it is important when a plane crashes and the cause of the crash may plausibly be due to a design defect (hardware, software, or their interaction). I mean that a design defect contributes to a reasonable degree, as virtually all crashes involve a confluence of multiple bad factors at once (see
Tenerife Airport Disaster, 1977, for a classic case of many bad factors at once).
In such a case, every plane of that type (or subtype) is usually at the same risk because the risk stems from a design defect.
In that case I think planes should be grounded until a reasonable determination can be made as to a design flaw and some reasonable mitigation applied.
There could be a short-term mitigation like pilot training on procedures or an MCAS shut off button added to the cockpit panel, followed by a long-term redesign like adding redundant AOA sensors and software to alarm the cockpit and disable MCAS when the redundant sensors indicate a discrepancy. After the Lion Air crash, Boeing has been working on a relevant software change and training change for pilots but I don't think either have been implemented, yet. Boeing has also been very justly criticized for not disclosing the existence and operation of these new systems in pilot training. That's not a good look for Boeing.
The reason for my thinking is not so much that the risk of a third aircraft crashing next week or next month is that high in absolute terms, but that such a potential crash would be preventable if the design defect is responsible and is not mitigated. Crashes that can reasonably be prevented should be prevented
both to preserve human life and to maintain the public's confidence in the safety of air travel.
To return to the Tenerife Disaster example, although many highly unusual factors contributed to the extreme congestion and very poor weather at a tiny airport, significant responsibility was ultimately laid on the radio communications and unclear use of language that led one 747 to take off while a second 747 was still on that runway shrouded in fog.
As a result, important and specific worldwide changes were made to radio communication protocols for take-off (quoting from the Wikipedia article):
Air traffic instruction should not be acknowledged solely with a colloquial phrase such as "OK" or even "Roger" (which simply means the last transmission was received), but with a readback of the key parts of the instruction, to show mutual understanding. The phrase "take off" is now spoken only when the actual takeoff clearance is given or when cancelling that same clearance (i.e. "cleared for take-off" or "cancel take-off clearance"). Up until that point, aircrew and controllers should use the phrase "departure" in its place, e.g. "ready for departure". Additionally, an ATC clearance given to an aircraft already lined-up on the runway must be prefixed with the instruction "hold position".
Even though the specific circumstances that caused Tenerife's loss of two 747s at once is extremely unlikely to ever be repeated in my opinion, the change above in communications has improved air safety for all flights, most especially when operations are far from routine at the airfield. In that case, the type and design of the aircraft was not a significant issue, but the protocol for communication was, and it was changed.
I hope this explains my thinking and I'm happy to discuss further if anyone is interested.