The use of RAI for remnant ablation, as adjuvant therapy or as treatment for recurrent or metastatic disease is undergoing a significant reappraisal, based on its questionable efficacy in various disease contexts. Until recently postoperative RAI treatment was given to all patients with thyroid cancer regardless of the pathological stage of the disease. Although its use for low risk forms of thyroid cancer has diminished, current indications for adjuvant RAI treatment are not based on randomized clinical trials, but instead on consensus expert recommendations that rely on retrospective studies, despite their methodological weaknesses and biases. Most thyroid cancers are driven by oncoproteins that activate MAPK signaling, and the transcriptional output of this pathway is inversely correlated with the expression of genes that govern many of the specialized functions of thyroid cells, including the ability to incorporate iodide into thyroid hormones. Most clinical guidelines recommend treating patients with tumors > 2 cm or with significant nodal disease with RAI after surgery. In our view this is problematic, as tumors giving rise to nodal disease are markedly enriched for BRAFV600E, which have a high MAPK pathway flux and are mostly unresponsive to RAI. We will discuss the genomic determinants of the thyroid differentiation state in thyroid cancer, their relationship to RAI avidity and the extent by which thyroid differentiation can be restored by treatment with selective RAF or MEK inhibitors, or with combinations of compounds that block the MAPK output more profoundly. We will also discuss insights arising from the analysis of patients with exceptional responses to radioactive iodine, and by contrast, the genetic lesions that determine irreversible loss of thyroid identity.