I think binding TAPs to key moments in time makes them unnecessarily fragile. I instead think of the triggers as affordances in the space of representations. E.g. ‘this isn’t an oreo sort of trip’ feels more available as a context setting representation at another relevant boundary: the inside-outside the store distinction.
Flexibility of representation is a double edged sword. It provides more affordance points and allows more transfer learning and modularity. It also gives you more outs and the chance to cold read-confirmation bias yourself into thinking wrong things instead of using the representations to cut down on the number of counterfactuals that you need to consider (a key feature).
I think binding TAPs to key moments in time makes them unnecessarily fragile. I instead think of the triggers as affordances in the space of representations. E.g. ‘this isn’t an oreo sort of trip’ feels more available as a context setting representation at another relevant boundary: the inside-outside the store distinction.
Flexibility of representation is a double edged sword. It provides more affordance points and allows more transfer learning and modularity. It also gives you more outs and the chance to cold read-confirmation bias yourself into thinking wrong things instead of using the representations to cut down on the number of counterfactuals that you need to consider (a key feature).