Our brain things at much faster pace than the pace at which the tools can capture the thoughts. We should choose tools that can capture the thoughts in real time but have enough scope for corrections too. We progressively add clarity to our thoughts and the tools also should be able to start taking the immediate historical context and do the corrections based on the newly available context.
Think of it like a mix of interpreter and compiler. Interpreter just interprets word by word so that we get the immediate output as to if the tool is capturing the thoughts or not. At the same time as the tool gets time to process and finds idle time(pauses by the speaker or there is more computing power than what is required to do the interpretation) then it starts compiling based on the surrounding context. It needs to look at the context of the current phrase, sentence, para and doc so that it can make much more sensible corrections to the grammar and sentence formation.
Today when using Google speech to text on Google docs and Otter I feel that this is already happening to a certain extent.
Referenced in:
All notes