You do present a valid flaw of a hypothetical one-time remotely possible transaction that may or may not exist.
He doesn't.
Even though the article is nice, I think that you're wasting your time with this thread. Initially, they were complaining that Segwit is complex and now they're making up scenarios in which it is supposed to be 'insecure'. I doubt that you're going to reach a point where they admit to being wrong.
While Segwit is complex and introduces many changes, it is still about the same number of lines of code as the Bitcoin Classic implementation of the 2 Mb hard fork because that implementation still needs additional changes to mitigate the problems with quadratic hashing.
This is false IMO.
On segnet4 consensus changes are in commits 7c68afbd747ad57391fcb66485c377298fb02a8e to 4dd3d7dd8bf2f9dd7a5e62c3cb2ca8dbd1146daa
Git diffstats says 65 files changed, 1262 insertions(+), 350 deletions(-)