Now that I’ve got my environment set up, I am ready to start experimenting with use case ideas. Previously I talked about creating a UX touch point on all my main devices that I can own and control, and connecting them all to a central service (that I also control.) That is now mostly done and I’m happy with the somewhat shake-y platform I have built and shared publicly.
There is a famous Bret Viktor talk that constantly shapes my thinking around UX: A thought / desire manifests in our heads, and then takes a certain amount of effort to enact into reality. When an app UX is good, there is a 1 to 1 to 1 mapping between thought, action, and result. When a UX is at the wrong level of abstraction it takes many sub-steps to go from what is desired internally to manifesting in reality.
Of course, some would say that in a perfect world we would be one with our actions and in a zone of no thought.
With this app I am setting up my first use case. Whenever I have a thought I want to act on in some way, I write it down here and start adding buttons to go the next steps. For example, I’m interested in translating phrases into multiple languages so I can try learn in parallel. Another example is that while I’m not sure how to act on all these thoughts yet, I can at least tag them and drop them in a bucket for later organizing. The workflow I’m envisioning involves some weekly review process in a more comfortable location. This will allow me to build an asynchronous workflow that doesn’t slow down my main path.
I am exploring two options for ‘end of week review’ at the moment, you can see them at the bottom of the left screenshot: week review & language retrospective.
Week review: At the end of the week, ChatGPT will take all the individual items (which contain notes such as ‘right now I’m buying a bagel’) and it will convert them into a first person narrative that can double as a journal entry. Instead of sitting down and writing the weekly journal entry, I write the microjournal entries and the weekly version is composited. On its own this is a net-negative, because one of the main benefits of journal entries is sitting down and writing them. BUT: I am also pairing this with ChatGPT review, where ChatGPT will pretend to be a psychologist or various famous historical figures while giving feedback and tracking progress.
Language review: The best learning people can do is for something they are intrinsically motivated for. It is much easier to learn how to open a cookie jar if you want a cookie. This weekly review will take all translation requests for the week and convert them into a cheesy 1 page story. I’m hoping this will be an easy way to review old phrases, and there’s no need to only include last weeks phrases. This can grow into an anki style spaced repetition algorithm.
As always, we’re very early to this new way of thinking (AI) so a lot of this will be experimentation. I will use this for a while and see how it fares in the field, before tweaking it further. I have a local dev environment which can pull these notes and do post processing, hopefully it comes up with something interesting.