There's a difference between looking at an average (median in this case) and saying that the average of a sampled population is then applicable to an individual, because it isn't, it's only a measure of the population.
Let me explain with an example:
If I were measuring shoe sizes in a school class, I would survey the children's feet (Bob might be a size 2, Jane a 4, Felicity a size 5, Derek a size 9 and Ranjit is a size 6.), I rank the data in order, 2, 4, 5, 6, 9, and the median value is size 5.
There's nothing wrong with the median shoe size, but if I need to get shoes for one of the children, there's a fair chance that they aren't a size 5, despite my average saying that it is. One size doesn't fit all, the average only describes the population as a whole, not the individuals.
The same for GI, no one is dismissing it as irrelevant, but it's based on the population surveyed, so probably not 100% applicable to each individual within that population, nor the public at large.
Testing yourself after particular foods is the sure fire way to find out if GIs published fit you or not. They might broadly fit you, exactly fit you, or not fit you at all.
You don't want to have to be the blood glucose equivalent of Derek, squeezing his size 9 feet into a size 5, or Bob, looking like he's wearing clown shoes and tripping over things because they are too large for him.