Manufacturing is super complex - employees and robots performing thousands of interconnected tasks, sometimes requiring a little bit of judgement, and never 100% supervised.
Even a mature operation like Toyota and Ford can blunder.
For me it was a situation where I've never heard of another case, so I guess highly anecdotal, but can see how it might have applied to millions of regular old Ford ICE cars that were on the road.
At one end of the accelerator linkage is the "user interface" (gas pedal) and at the other end it's the entire V8 engine. Where the linkage connects directly to the throttle using optimized leverage and failure-mitigating springs that can overcome a number of foreseeable failure scenarios. Which had gradually been improved since the Model T and through the entire Space Age.
Runaway acceleration wasn't a problem with the pedal.
Engine was running properly too. And the linkage was perfect.
Well the engine is heavy and is not bolted directly to the frame of the car, instead it uses motor mounts, which consist of a metal plate which bolts to the motor, and an opposite plate that bolts to the frame, separated by a thick hard rubber shock absorbing pad.
Motor mounts are doing some of their isolation duty when you see an idling engine under the hood shimmying a little while the fenders and hood are nicely stationary.
Anyway it took years to figure out because it happened so seldom, and the cause & effect were so widely separated in time, but one day while navigating a jeep trail the engine had bottomed out, and that must have been the time one of the motor mounts separated, while the engine was momentarily forced an inch or more away from its normal position.
Nothing ever seemed any different, but every once in while when I would accelerate from a stop, the pedal would drop almost to the floor and it would really take off until I got my foot under the pedal and pulled it back.
Turns out the torque was occasionally capable of twisting the engine in the direction away from the broken mount, enough to be pulling on the linkage, opening the throttle further, and resulting in more torque. Might have had something to do with the octane of the gasoline in use.
I guess at either end of the linkage you want the rubber to be bonded to the metal a lot better than you might think at first.
Even a mature operation like Toyota and Ford can blunder.