Since the realization that the cellular homologs of a gene found in the retrovirus that contributes to erythroblastosis in birds (v-erbA), i.e. the proto-oncogene c-erbA encodes the nuclear receptors for thyroid hormones (THs), most of the interest for THs focalized on their ability to control gene transcription. It was found, indeed, that, by regulating gene expression in many tissues, these hormones could mediate critical events both in development and in adult organisms. Among their effects, much attention was given to their ability to increase energy expenditure, and they were early proposed as anti-obesity drugs. However, their clinical use has been strongly challenged by the concomitant onset of toxic effects, especially on the heart. Notably, it has been clearly demonstrated that, besides their direct action on transcription (genomic effects), THs also have non-genomic effects, mediated by cell membrane and/or mitochondrial binding sites, and sometimes triggered by their endogenous catabolites. Among these latter molecules, 3,5-diiodo-L-thyronine (3,5-T2) has been attracting increasing interest because some of its metabolic effects are similar to those induced by T3, but it seems to be safer. The main target of 3,5-T2 appears to be the mitochondria, and it has been hypothesized that, by acting mainly on mitochondrial function and oxidative stress, 3,5-T2 might prevent and revert tissue damages and hepatic steatosis induced by a hyper-lipid diet, while concomitantly reducing the circulating levels of low density lipoproteins (LDL) and triglycerides. Besides a summary concerning general metabolism of THs, as well as their genomic and non-genomic effects, herein we will discuss resistance to THs and the possible mechanisms of action of 3,5-T2, also in relation to its possible clinical use as a drug.