²⁴Understanding why and how the thermostat suffers creates an interesting illustration of applying the theoretical ideas of this chapter and the preceding chapters. Let us assume that the thermostat is predicting a frustration of its main goal to be likely to happen in the future. The theoretical problem here is that there is no frustration yet, so it is not obvious how this would lead to suffering now, if we merely rely on the basic frustration and reward loss theories of Chapters 3 and 5; we need to consider something more. First of all, the interpretation of RPE as being able to compute frustration based on comparing two predictions as in footnote 23 above (or footnote 20 in Chapter 5) is one option: when the thermostat reads the negative weather forecast, its prediction of total future reward will change for the worse, and this triggers frustration. However, in this chapter we found several other explanations for that phenomenon. If the thermostat realizes it is unable to properly control the temperature in the future, the uncontrollability may trigger a negative internal reward, and a reward loss. If this happens often, it could also be that the self-evaluation system of the thermostat concludes that it is not performing its central task well enough, thus leading to frustration due to the self-evaluation. While it may be somewhat contrived to apply the theory of survival mechanisms or threats to the self in this simple example, it is possible that if the thermostat fails to keep the temperature constant, it will be thrown into the garbage bin, and a programmer might decide to explicitly program self-preservation mechanisms that produce a negative internal reward when such “death” seems to be approaching—this is the mechanism of suffering due to threat-based internal reward. Actually, if the thermostat is really hyperintelligent, it might figure out, by itself, this possibility of being plugged out, and try to behave accordingly, thus creating a self-preservation mechanism on its own. There is yet another possible mechanism, not treated in this chapter but in Chapter 9: the prediction of future difficulties could take the form of a simulation that triggers the frustration mechanism based on the frustration present in that simulation. Thus, we see many mechanisms that may create suffering in the future-looking thermostat. Whether these frustration-based mechanisms really corresponds to what one would usually call fear is a complex question that I cannot conclusively answer at this stage.