Two key facts:

1. Antecedent didn’t occur (much less than 40% unvaccinated).
   CDC data summarized in August 2023 reporting shows that by May 11, 2023, about 81.4% of the U.S. population had received at least one COVID-19 vaccine dose, i.e., only ~18.6% remained entirely unvaccinated. (frequentbusinesstraveler.com) Later reporting notes that by 2025 about 70% of Americans had completed the initial vaccine series, implying ~30% were unvaccinated or incompletely vaccinated—still well below 40%. (nypost.com) So the condition “if ~40% of Americans remain unvaccinated 2–3 years from mid‑2021” was not met.

2. No new “killer strain” emerged in 2023–2024.
   The dominant variants in that period were Omicron sublineages such as JN.1, KP.3, LB.1, and later NB.1.8.1. Multiple analyses from WHO, CDC, and academic/medical sources consistently report that these subvariants are more transmissible and more immune‑evasive but do not cause more severe disease or higher mortality than prior Omicron strains:

   • JN.1: WHO and CDC data show no evidence of increased severity vs other Omicron variants. (yalemedicine.org)
   • KP.3: Described as more contagious but not associated with more severe illness than recent strains. (health.com)
   • LB.1 and NB.1.8.1: Monitored variants with no evidence of more severe disease or higher hospitalization/death rates than contemporaneous variants. (cbsnews.com)
     Broader reviews also note that post‑Omicron, COVID‑19 has generally become more transmissible but less severe, with lower case‑fatality rates compared with earlier waves like Delta. (lemonde.fr) There was no distinct “highly lethal” new U.S. variant in 2023–2024 that caused catastrophic mortality specifically among the unvaccinated.

Why this is scored as “ambiguous”:
Chamath made a conditional forecast: if ~40% of Americans stayed unvaccinated 2–3 years out, then a “killer strain” would emerge with essentially 100% probability. In reality, the vaccination rate rose high enough that this 40%‑unvaccinated scenario never happened, and pandemic evolution took place under different immunity conditions. Because the stated condition failed, the prediction’s implied causal claim about what would have happened under that counterfactual cannot be directly tested.

If one ignores the conditional and treats his statement as an unconditional claim that a killer strain would appear by ~2023–2024, available evidence would make it wrong; but under the literal conditional phrasing, the fairest scoring is that the prediction’s truth cannot be determined from realized data, hence “ambiguous.”