Aldose reductase (AR) enzymatically transforms cytosolic glucose into sorbitol, a molecule that poorly penetrates cell membranes and is sometimes slowly metabolized. Hyperglycemia can cause intracellular accumulation of sorbitol and its metabolite, fructose, which can create osmotic swelling and cell dysfunction. Driven by this simple paradigm, the "Osmotic Hypothesis," and armed with positive pre-clinical results on prototype AR inhibitors (ARIs), researchers worldwide have targeted diabetic neuropathy with ARIs for four decades. However, most double-blind placebo-controlled ARI diabetic neuropathy trial outcomes have been disappointing. Ironically, scientific evidence that AR plays a key pathogenic role in diabetic neuropathy has continued to mount. Diabetic mice lacking AR exhibit strong protection of nerve function. Diabetic mice overexpressing AR have accelerated nerve dysfunction and damage. Human diabetics with "high AR expression" alleles shows faster loss of maximum pupillary constriction velocity, an indicator of autonomic neuropathy, while those with "low AR expression" alleles have slower loss of foot hot thermal threshold, an indicator of sensory neuropathy. Evidence is now strong that the Osmotic Hypothesis and the nerve sorbitol endpoint were misleading. Reliance on nerve sorbitol to assess AR inhibition likely caused underestimation of doses needed for clinical efficacy and overestimation of drug safety margins. Current recognition of the pathogenic importance of oxidative stress and its strong link to metabolic flux through AR have led to a revitalized "Metabolic Flux Hypothesis" emphasizing cofactor turnover rather than polyol accumulation. Hopefully, these new insights will lead to novel ARIs that will effectively and safely slow the progression of diabetic neuropathy.