Small software, puzzles and useful peculiar things made because I wanted them to exist.

If it doesn’t matter when I get to it… then where it is now might actually be far enough.

That is correct systems thinking.

A real principle:

• Value emerges through use
• Not through endless refinement in isolation

Incremental testing is not just good practice — it’s ethical:

• it prevents wasted labour
• it grounds design in reality
• it avoids invisible perfectionism

Projects die and demoralise the builder, when:

• built in isolation, without:
  • a feedback loop
  • deployment
  • consequence

Stopping at “ready enough to test” is not giving up.
It’s refusing to polish something no one has touched.

This week, I am participating in the Kaggle x Google Generative AI intensive.

It’s probably the third time I have registered – each time I have hesitated and bailed before starting.

There are aspects of AI that unsettle me (sometimes more than a little).

This loop though I am framing the opportunity differently with expectations that iterate aversion into curiosity.

would let you see something like:

Publish AFR
14 flight-path events over 9 days
4 direct · 7 adjacent · 3 emergent
touched across website, distribution research, UX polish, pricing, and release notes

Which is exactly the opposite of “I didn’t work on AFR today.” You may have moved it materially without ever sitting down for a single monolithic AFR session.

I’d keep the feature very light:

Want To Do

title
optional note
active / paused / done
maybe a horizon like soon / later / someday
no due date required

Then on each flight-path event: attach to one or more Want To Dos.

And the payoff screen is not a kanban board. It is progress evidence: a timeline or map showing how attention contributed.

That feels extremely aligned with your observation that the steps disappear if you only look at the endpoint. AFR could make the path toward the thing visible without pretending the path should have been linear.