Because Twitter and Mastodon are interchangeable here. If this same person posted the original content as a Tweet and there were a pattern of users posting Twitter links to boost Twitter engagement, the problem would stand.
I would note as well that the Mastodon instance clearly struggled with the initial load from the HN traffic. Testing its load time when few users from HN were being directed to it won't be a very good benchmark against Twitter.