The wikipedia article for Original Antigenic Sin (OAS) oversells it a bit, in my subjective analysis. While the effect is experimentally verified - e.g. patients with a Dengue type 4 infection will produce antibodies that bind type 1 much more strongly than 4 if they were first infected with type 1 (source) - there is no evidence I'm aware of that proves the resulting immune response is less effective than a naive immune system of the same age would mount. From Original Antigenic Sin: How First Exposure Shapes Lifelong Anti–Influenza Virus Immune Responses:
Although OAS has often historically been depicted as a problematic response, recent data have demonstrated that, in certain contexts, eliciting OAS may also be beneficial.
It's not obvious that the poor match of OAS antibodies to the second infection is enough to make up for the much more rapid response compared to that of a naive immune system. Mismatched antibodies generally still neutralize other strains, it's just that a higher concentration is required.
A big confounding factor here is another mechanism, Antibody-Dependent Enhancement. ADE is most infamous in Dengue, which causes mild disease the first time but can be deadly after that if later infections come from a different strain than the first virus. ADE arises in this case because Dengue virus can infect the white blood cells that envelop and attempt to destroy antibody-bound particles. This results in accelerated virus production and a dysfunctional immune response. Despite the fact that immune cells do not display the ACE2 receptor that their spike protein needs to bind before it can fuse with the cell membrane, both SARS and SARS2 have been found to also infect macrophages when spike-binding antibodies are bound. Unlike Dengue, though:
Macrophages infected with SARS-CoV, however, did not support productive replication of the virus.
SARS-CoV-2 infection thus produces antibodies that elicit ADE of infection, but these antibodies do not contribute to excess cytokine production by macrophages.
Which, together with the protective effect observed for SARS2 vaccination and prior infection, we can conclude ADE isn't an issue for COVID severity, even if variations arise that can escape antibody neutralization.
Finally, there is some important context here that is not obvious to the casual investigator. The new omicron variant of SARS2 has been described as "highly mutated" because it has more than 30 mutations in the spike protein, out of 1250 amino acids in the full sequence. In other words, the other ~97.6% AA are identical. For comparison, I looked up and ran a BLASTp analysis of hemaglutinin (analogous receptor binding protein in influenza virus) from H1N1 and H3Nx strains of influenza. (uniprot accession nos. P03452 and F1KLX3, respectively) These two antigens have only 42% identical amino acids. So they are far more dissimilar than different variants of SARS2. COVID is not likely to reach similar levels of differentiation anytime soon, and we can expect any potential OAS effects to be mild for a long time as well.