Uhm, that's not how statistics works.
You can't take a few selected examples and generalise to the entire population.
There are all sorts of sampling biases in this. For one, all the samples are people trying to prove, or were surprised to find that in their case, EV batteries are reliable. They're also EV early adopters (by definition), and so have a stronger motivation to report success.
Also, we have no data on failures, people whose EV batteries died unexpectedly quickly.
A meaningful analysis would be to get data from a wider set of randomly selected owners (preferably > 1,000) across multiple makers for a period of a few years and see what the longevity rates look like across the board.