There is a moment in almost every new workflow when the novelty wears off. The tool can do a hundred things, but what is the one thing I actually need it to do?
That is the question I keep coming back to. The most interesting uses of AI are often the least theatrical: making a messy note useful, spotting the next action in a conversation, or helping me get an idea onto the page.
Start with the friction
Instead of asking what a new tool can do, I want to notice what gets in the way. Which jobs do I keep postponing? Where do I copy the same information twice? Which good ideas disappear because recording them takes too long?
A useful experiment starts there. Pick one small piece of friction. Decide what a better result would look like. Then see whether the tool can help.
Keep the human part
An assistant can prepare a draft, but the judgement still matters. Does this sound like something I would say? Is the example accurate? Would it actually help someone?
For this journal, the intention is simple: make it easier to collect ideas, and leave enough room to think about them before they become public.
The interesting question is not how much we can automate. It is what we want to make more space for.