The internet laughed at Apple this weekend due to their terrible progress in AI and the irony of publishing research taking on today's reasoning models.
However, the paper is quite insightful and rigorously highlights the limits of today's frontier reasoning model capabilities.