A small pile of questions to resolve here, that are important in considering macros timelines.

1. There is a possibility that we launch a single macro, e.g. the JSON macro, without opening up the macro feature for general use. If we do that, what state does the augmentation feature need to be in? Is it okay to ship like that if e.g. augmentation feature spec is still in progress? Implementation? Tests? Do we need to split out a distinct feature?

2. There is a possibility that we launch a less powerful macro feature, such as the "phase 3 macros" recently under discussion. If we do that, what state does the augmentation feature need to be in? Do we need to split out a distinct feature?

3. There is an argument for launching augmentations before macros to reduce the size of the macros launch. Do we want that? Should we anyway do the work to separate out augmentations as a distinct feature--currently it's the "macros" experiment flag.

Discussing this probably fits naturally into macro timeline discussions, I'm filing mostly so it shows up as a blocker for having phase 3 macros ready for launch.