I did not have a clear plan in mind. I am in a phase where I am spending more time exploring, for different reasons. One of them is paying more attention to where the world is heading, what I want in life, and building the right mental models.
We still need to keep acting. On my side, I still need to optimize how I test an idea. It is a shame to lose inspiration, to miss the chance to test something right away, or to skip building that useful tool you need, want, and could have today only because you do not have the right pipeline.
I tested many small things, rebuilding from scratch each time and trying different services and stacks. I have also spent a lot of time learning, trying to master things mostly through study and only applying what I learned when I could. That habit needs to change too. I want to approach learning and doing differently.
Ask hard questions. Most choices sit on top of assumptions and requirements. I need to put those requirements and assumptions back on the table more often. If I do not, the world changes and I stay blind to it while I keep steering by an old map.
So I try to keep this prompt ready: weigh pros and cons, consider better alternatives, notice blind spots that any model leaves out, and then apply your own judgment. Sometimes the old assumptions are still right. Often they are not.
And here we are, in the game of life. The first fork in the road:
- Do I keep compounding on things that might be useful on the career path I have followed (whether I meant to or not)?
- Or, if instinct or circumstances push me toward something completely new, do I go that way?
Validate first. How can I validate this idea before I over-invest in building it?
Derek Sivers makes a related point in How to decide and make the best choice. I am paraphrasing the spirit, not quoting him: when there are too many options, there is often no single “absolutely right” choice sitting there waiting. You pick one, and that choice is the right one until new information says it is not. The move matters more than endless comparison.
So the practical questions for me are: what to build today, which approach, and which requirements matter today (they may not be valid tomorrow anymore).
Another requirement: I want to understand how and where to use AI and LLMs in a way that actually makes sense. A lot is shifting in software engineering and data engineering when coding assistants keep getting better. I want a clear view of what that changes for how I work, and what it does not change.