I want to do work I can take seriously: intellectually, practically, and in what it does for other people. I am ambitious. I am drawn to difficult problems and to tools that expand what a person can do. I also want a better answer to “Why this?” than that it was technically possible or looked impressive.
Helping someone understand a difficult subject or making an ordinary task less frustrating can matter on its own. The people affected by the work should count in deciding what is worth doing, as well as in evaluating the finished result.
Useful to whom?
I gave Iron Protocol free to younger and older people who were new to exercise and healthy eating. Making something available is one part of helping; making it usable is another. I want someone to leave with a clearer next step and the confidence to approach it.
People bring their own goals, knowledge, constraints, and preferences. Those circumstances help determine what a useful solution should be. A technically sound answer can still be difficult to understand or impractical to follow. I want the person receiving it to be able to do something with it.
An answer has to earn trust
At QDA, our team began incorporating AI and natural language processing into its algorithms in 2018. More recently, I have used AI for nonprofit analysis and tools I design and use. The possibilities interest me, but so does the question of what justifies relying on the result.
In nonprofit research with FCCF, I challenged questionable classifications and record matches and requested source verification. A record that looks complete can still contain an unresolved assumption. Someone using it to make a decision needs to know that.
Iron Protocol’s written rules make a related distinction: planned exercise must not be recorded as completed without the user reporting it. These rules define intended behavior, not a guarantee of perfect execution. The standard I want to hold the work to is that someone relying on it can distinguish what is supported from what is assumed.
Finding a direction
My path through founding and leading QDA, nonprofit work, and fitness was not a sequence I planned in advance. Those experiences remain relevant without having to be steps in one long-term strategy.
Creating a useful system means understanding how its people, processes, and tools affect one another. Improving it means finding where those relationships break down and checking whether a change solves the problem. That is the work I want to do in AI systems operations.
Put the idea to work
Scrutiny should lead to better decisions. Sometimes an unresolved assumption needs more investigation. Sometimes a limited experiment is the honest next step, with clear boundaries and a clear account of what it can and cannot establish.
I want to build ambitious things and remain willing to change them when someone else’s experience shows me that I have solved the wrong problem. The people using the work should have a say in what counts as success.