Again this is just wrong.

If you've been vaccinated then you have B-cells that match Omicron, even though you weren't vaccinated with Omicron's spike.

Maybe my analogy didn't work, but you have a flawed understanding of the immune system.

As a concrete and well-studied example, the H1N1 strain of influenza stopped spreading in the human race in 1957. Nobody after that was exposed to it, everyone born before that was exposed going back to 1918. It was displaced by the H2N2 pandemic in 1957. It mutated in pigs for 50 years. Then it underwent triple recombination in 2009 and jumped back to humans -- with an H1 gene that was derived from the human strain and an avian N1 gene that hadn't ever been seen before.

The impact of that 2009 pandemic was considerably blunted by human cross reactive T-cell immunity in the older population who had seen the long-lost relative of H1 protein back before 1957. That 50 year old "pre-vaccination" did not wane and was cross reactive after 50 years of mutation and adaptation in pigs, and was still effective enough to protect against severe disease and death.

Nobody had immunity against infection of course, but that cross reactive immunity to H1N1 in all the older people helped to dramatically blunt the impact of the 2009 pandemic to the point where the human impact was lower than normal human influenza.

Cross-reactive B-cells and T-cells to future variants to SARS-CoV-2 are going to work the same way. Most of the epitopes will still be conserved and the T-cell reaction will be able to clear the disease and protect against severe disease and death.

You don't need reimmunization every time the virus mutates slightly, that would be poor design. The humoral immune system works closer to a Bayesian spam filter (arguably better).