The Airplane That Does Not Fly w/o a Computer

This is tough one. When I read about a military stealth aircraft, the feelings were mixed at best: it is a super-expensive over-complicated aircraft, which did not have one feature the aircraft should have: ability to fly well. The reason was is that the shape of the aircraft was disctated by the need for it to remain stealthy, which was more important that it's air-dynamic and air-gliding abilities. So the only way for this thing to be stable in the air is to have very powerful computers corresting it's position in the air many times as second. What could possibly go wrong in this design???

The trade-off was well-justified at the time - but this approach never made me feel at ease. This approach really hit the fan when a large commercial airline manufacturer applied this concept to its passenger jet. The engines installed onto the existing body were too large for the body, and because they were so large they also have to be moved up and forward. This design was driven by the business requirement from the airlines, asking for a plane which could carry more (and more... and more...) people inside the same (more or less) narrow-body jet. So such and "intersting design" created some "interesting" flying dynamics: it requires to have a computer frequently correcting angle of the flight during take off. Which later lead to a few very-very tragic outcomes...

Hard to spot and could have many shapes and forms. I usually apply this to an opininated design, which is practically impossible to change because of large number of proprietory components. A good example is a prorprietory cloud provider design lock-in.

What To Do Instead

This large expensive problem has a very simple fix. I often see people offer management a slightly more expensive "generic" or more portable design, which mean this "air plane" can fly on it's own, and then a slighttly cheaper proprietory design - but that fits the proprietory architecture dictated by your cloud provider. The management (or less technicaly leadership) does not even blink: they have been convinced by the cloud provider's many articles and videos, and conferences, that the only effective way to do things - is the proprietary way the cloud provider is pushing so very hard to adoption. So they chose the proprietory approach without even giving it any second thought, thus creating a product (the airplain) which cannot fly without its proverbial "computer" (the cloud).