Usually everyone does much better in rapid than bullet because more time to think generally means higher quality games. Who are "so many people" with 1200 rapid and 1800 bullet? The few who this is true for probably focus on bullet and never or seldom play rapid, but I can assure you that if someone has 1800 bullet, then I'd expect their rapid rating to be in the ballpark of 1700-2000 given a decent sample size of games.
Check their profile stats to see how many games they played under those time controls - not just the ratings listed
How come in rapid i do so much better than in bullet even though i see so many people with a 1200 in rapid and a 1800 in bullet?
Bullet: https://www.chess.com/a/2BGC4ncJzjuE6
Rapid: https://www.chess.com/a/39MFkq4GNjuE6
In rapid i had 99.0 accuracy and in bullet i had 13.8 accuracy.