I’m building a personal agent system for myself. I want it to understand what I’m working towards, hold the relevant context and carry useful work forward without me having to drive every individual step. The 100× ambition is where I’m aiming. I’m interested in finding out what it would actually take to get there.
My work moves across product, strategy, brand, design, writing and the practical business of turning ideas into something real. I like being involved in all of it. That is also why this interests me so much: there are only so many things I can personally give proper attention to in a day, however much I want to do.
A lot of the opportunity sits in the work around the decisions. Researching an idea, finding the right information, preparing a brief, comparing approaches, pulling a document together and carrying a decision into the next piece of work. Each part takes time. So does remembering where everything stands when I come back to it.
I want a system that helps me carry those threads. If I’m exploring a product idea, it should understand why I’m interested, what I’ve already considered and which assumptions need examining. It could research several approaches while I’m doing something else, prepare the useful differences and bring me back a decision I’m ready to make.
Once I make that decision, I want the work to continue. The reasoning should carry into the plan, the brief and whatever we need to create. I shouldn’t have to explain the whole thing again because I’ve moved from research into writing, or from a conversation into a design tool. The work should keep its context as it moves.
That is the part I find exciting about building something personal. It can be organised around the way I think and the standard I expect. Over time, I want it to retain what I’ve corrected, what I’ve rejected and what has actually helped. If I keep having to give the same feedback, then I haven’t really built the system I have in mind.
I’m also curious about what happens when it can do useful thinking between the moments I’m actively using it. Something changes in an area I’m following, and it connects that change to a project I’m considering. It could investigate whether an assumption still holds and bring me the implications, rather than leaving me to spot everything myself while trying to keep up with the news.
I don’t expect every task to become 100 times faster, and I’m not claiming I’ve achieved that improvement. Some work will still depend heavily on my judgment. But I do think the combination of parallel work, remembered context and less coordination could create a much bigger change than speeding up a few isolated tasks. That is what I want to explore.
The test has to include my time. If I get a hundred outputs back and spend the evening checking them, I haven’t gained what I wanted. I care about work reaching a useful standard, how much of my attention it took and whether the result helps me make progress. More activity on its own would be a fairly disappointing outcome.
I want to write about this as I build it: the tools I choose, the parts that work, the things I have to change and the results I can actually stand behind. I have plenty of ideas I want to give a proper attempt. Finding out how much more I can do with a system built around me feels like a good place to start.