I'm not one that's completely for that type of food reward hypothesis of obesity, but it is a real thing.. I just don't think it's significant enough in a majority of the population.
High sugar foods definitely stimulate endogenous opioid synthesis, which increases dopamine signaling through the mesolimbic system leading to reinforcement of the behavior aka an "addiction" to eat, though I think that's where the overstatement begins. Notwithstanding, mu opioid receptor stimulation also increases food intake/hunger and increases palatability of foods. There's direct interaction between the NAc for instance, and regions of the hypothalamus that regulate satiety/hunger as well.
Again, I do think that's overstated overall.. I think this may actually happen for a small group of the population though. I'm sure simple reinforcement may play a quite subtle, but large role overall.
Nonetheless, if they really want to delve away from a thermodynamic model of obesity, other avenues are imo at least more likely & significant, such as chronobiology, the gut microbiome, inflammation, leptin resistance, etc. Though these are still secondary to thermodynamics. Honestly they could probably all amalgamate and figure things out much better.
In general, it would go in a a fashion somewhat like initially a slight reinforcement in food intake via highly palatable foods, leading to slow fat accretion --> alters gut microbiome and alters intestinal permeability allowing more LPS to pass through, causing systemic inflammation --> stimulates a lot more leptin, chronically, leading to initial vagal afferent leptin resistance --> cross talk with other pathways causes increased resistance to other satiety hormones acting at the vagal afferents, causing further decreased satiety and more hunger --> hypothalamic leptin resistance sets in, which also causes reciprocal insulin resistance via signal transduction cross-talk --> you're fucked
Chronobiology could fit in any category to speed things up, depending on the person's habits.