A recent article examines the practical capabilities and limitations of the formal specification language TLA+ for system verification. The author pushes back against the hype suggesting the tool can single-handedly solve complex AI safety challenges, clarifying exactly what it can realistically check and where its boundaries lie.