If this model is true, then that would seem to support the effectiveness of So8res' "goal hacking" strategy. If you can convince yourself that working towards the instrumental goal (eg. of "having worked out today") is a want-to task, then you would feel less depleted and could either do it longer or be better able to perform your next have-to task. Or, if you're performing these kinds of tasks after work, convincing yourself that they are "want-to" tasks could help you complete them even in your "have-to depleted" state.

However, I still can't quite shake the intuition that there is something being depleted just by doing "hard" mental work, even if it feels like a "want-to" task. Though, this proposed model does help explain how I can sometimes come home from a full day of work and get absorbed in a programming project for another 6 hours.

Thanks for posting these, Kaj. If the standard model of resource depletion is in fact wrong, then this is very important information for people to base their anti-akrasia strategies on.