The Mount Sinai team's JACS study combined AlphaFold2 predictions with X-ray crystallography to uncover a cryptic allosteric pocket on PKMYT1 missed by AI, virtual screening, and follow-up tools including AlphaFold3 and Boltz-2. This was an in-vitro structural and biochemical investigation, not an RCT, with limited sample throughput typical of crystallography workflows and potential conflicts from authors leading the institution's AI drug-discovery center. The work reveals kinase conformational plasticity that static models routinely underestimate, echoing patterns seen in earlier kinase inhibitor programs where ATP-site overlap caused off-target toxicity. A 2021 Nature paper on AlphaFold demonstrated strong performance on rigid folds yet flagged dynamics as a core weakness; a 2023 Cell Chemical Biology analysis of WEE1/PKMYT1 dual inhibitors similarly showed that minute ligand changes flip binding modes, confirming experimental validation as non-negotiable. Together these sources indicate that AI accelerates hypothesis generation but cannot replace biophysical confirmation when selectivity is the therapeutic goal.