The last pair of quotes feed into an idea I've been wondering about: is sunk cost bias a useful bias? That is, it seems to me humans frequently underinvest in long-term actions; if we do so, and we are often under the control of desires or blood sugar problems, then it would make sense to continue with long-term projects - 'engage in the sunk cost bias' - if we choose to start the project while rich in willpower but then have to continue the project under less optimal conditions. A good heuristic would be to be stubborn and continue projects even if they seem bad at that moment.

Two other bits of the article are interesting in the light of sunk cost bias. Sunk cost bias, oddly enough, does not seem to appear outside humans/primates. The article points out that choosing trade-offs are the hardest decisions and ones most often engaged in by humans and the very first kind of decisions to degrade in humans; and also that the sugar seems to act not by increasing total sugar but by affecting whether short-term or long-term decision-making areas are activated. This suggests a possibility: the normal brain decision-making structures begin to break down in humans for fundamental metabolic reasons, leading to suboptimal decisions... in the absence of a 'bias' in the opposite direction.

(Baroque or adhoc? Maybe. On the other hand, a lot of mental and biological processes seem to be regulated just by making an opposing process, rather than simply reducing the over-active process.)