But situations are never "the same". The input (camera images, sensor data) will always be (slightly) different and – in this example – might just lift the conf

But situations are never "the same". The input (camera images, sensor data) will always be (slightly) different and – in this example – might just lift the confidence of "this is a truck standing on the road" over the threshold.