Enhancing TalkTastic: Features for Improved Editing and Contextual Learning

The ability to configure a list of common names or acronyms. This is because I'll frequently reference executives (or my wife, whose name has one of three common spellings) who aren't in the explicit context, and I always have to go back and edit the spelling. Talktastic should be able to let me specify or learn nouns from other contexts and save them. Also important since I'm in a very technical field, and there's just a lot of acronym soup and terms that aren't in the explicit context that TalkTastic can't otherwise know.

The ability to set some simple context, and tell TalkTastic up top what context to use. i.e. hit F5, say "this is a note of appreciation to the team" or "I'm thinking out loud" as these need wildly different tones/formatting, which aren't obvious from the context, but can create significant rework of what TalkTastic otherwise puts out.

Related to both, the ability to spit out two paragraphs that I can choose from, when TalkTastic realized that I'm in a technical context and it's not sure how to best transcribe. For instance, if I'm going down a technical rabbit hole in a response, nearly always TalkTastic cleans up the details and the critically important stuff I talked through is missing, and then I have to scrub super carefully, or go back to the transcript and merge things. I'd almost prefer it to spit out a "I think this is right, but here's exactly what you said, in case I got something wrong." Better still would be the way to paste my corrections back to TalkTastic, so it could learn (I frequently do this when working with ChatGPT directly-- I give it my edits back, and ask it to summarize the edits and tone/language tweaks, and save those for future work).

And following up on that point: since I've started using TalkTastic more, I've used ChatGPT less, and it feels like there's less explicit learning because there's no feedback loop on the final tweaks I made back into the system. Some way to tell TalkTastic to suck down the final version sent/posted, would be amazing, even better with a question or two. "Oh, you seem to use a more friendly tone. Do you want me to do this for you next time?" or "In Slack, you always use emojis to start paragraphs and always tag at least one person with a note of appreciation at the end, do you want me to start working this in for future slack messages?"