The virus has a fairly low mortality rate for healthy people, most people aren't healthy.
The fact the vaccine is new is irrelevant because generally speaking you can substitute time with quantity. Most effects are normally distributed in time, and so if you have enough doses given, you can replace the fact it hasn't been around too long with a high number of administrations when making a safety assessment. And it's been given literally billions of times. We've seen the full gamut of consequences.
But anyways, 'long-term side effects' don't refer to latent effects. It's not things that magically appear 10 years later. They're effects that last a long time but usually onset almost immediately. There's no reason to think that 5 years of data is better than 3 years of data when administered billions of times.
And frankly if you think 3 years and billions of data points are insufficient nothing's going to change your mind. So you may as well stop pretending it's a data issue.
As I said, irrationality is your prerogative. But you are wrong.