According to Lavoisier, ‘Life is combustion’. But to what extent humans adapt to changes in food intake through adaptive thermogenesis – by turning down the rate of heat production during energy deficit (so as to conserve energy) or turning it up during overnutrition (so as to dissipate excess calories) – has been one of the most controversial issues in nutritional sciences over the past 100 years. The debate nowadays is not whether adaptive thermogenesis exists or not, but rather about its quantitative importance in weight homoeostasis and its clinical relevance to the pathogenesis and management of obesity. Such uncertainties are likely to persist in the foreseeable future primarily because of limitations to unobtrusively measure changes in energy expenditure and body composition with high enough accuracy and precision, particularly when even small inter-individual variations in thermogenesis can, in dynamic systems and over the long term, be important in the determining weight maintenance in some and obesity and weight regain in others. This paper reviews the considerable body of evidence, albeit fragmentary, suggesting the existence of quantitatively important adaptive thermogenesis in several compartments of energy expenditure in response to altered food intake. It then discusses the various limitations that lead to over- or underestimations in its assessment, including definitional and semantics, technical and methodological, analytical and statistical. While the role of adaptive thermogenesis in human weight regulation is likely to remain more a concept than a strictly ‘quantifiable’ entity in the foreseeable future, the evolution of this concept continues to fuel exciting hypothesis-driven mechanistic research which contributes to advance knowledge in human metabolism and which is bound to result in improved strategies for the management of a healthy body weight.