Refraction problem with globe model

We’ve all heard that when the sun sets, by the time you’re seeing it over the horizon, it’s already behind the alleged curve due to refraction. See the following image: https://imgs.search.brave.com/xiHNJMHIdwykdYtFVfYeNYAMPEBpxJ_BqjdjOT3vPzY/rs:fit:750:500:1/g:ce/aHR0cHM6Ly9jLnRh/ZHN0LmNvbS9nZngv/NzUweDUwMC9hdG1v/c3BoZXJpYy1yZWZy/YWN0aW9uLnBuZz8x If this were true, then the speed at which the sun travels toward the horizon (from our perspective) would drastically slow at some point just above the horizon, as the true sun continued to dip below said horizon, and the image we see would remain in position while refraction occurred. Then the sun we see would continue to sink slowly until it disappears. This doesn’t happen. The sun sets at the same constant speed at which it just traversed the sky. https://m.youtube.com/watch?v=ylLqqI7A1Gg