Recent deep learningDeep learning methods have shown high accuracy to predict neuro-ophthalmic and neurological disease, based on analysis of retinal images alone. Despite these important recent discoveries, more prospective, real-life studies are needed to validate DL algorithms as potential diagnostic aids in relevant clinical settings. Emerging Large language modelsLarge language models have shown variable performance in the subspecialized area of neuro-ophthalmologyOphthalmology, requiring more validation prior to their integration into daily practice.