Not arguing it doesn't. Just that using thermodynamics to understand human metabolism is akin to trying to grok flight through Newtonian mechanics. You can do it. But mostly by working backwards from higher-level principles.
The evidence for that disconnect is this thread. CACO leads people to believe you can't decrease your calorie consumed and burned without also increasing physical activity levels. Yet that's empirically incorrect. Not because thermodyanmics doesn't apply. But that it applies in mechanisms so complex that starting from first principles doesn't work until you have the roadmap.
> slight overeating for one body weight just means that you reach equilibrium at a slightly higher body weight
Studies observing the effects of calorie restriction see improvements in a variety of biomarkers. That slightly-higher weight comes with higher inflammation, higher CVD frequencies and other nasties.