Project Deck started with my 96-year-old grandmother's favorite card game.
The point was not only to build a digital version of the rules.
The point was to make interacting with the game feel like working with a real deck of cards.
That difference matters.
The rule is not the whole game
A card game is not just a state machine.
It is the deck. The dealing. The rhythm. The small physical cues that tell you what is happening before anyone explains it. If software keeps the rules but loses the feeling, it may technically be the same game, but it no longer feels like the same game.
That was the challenge.
Could agents help build something specific enough, tactile enough, and careful enough to preserve the feeling?
Project Deck / ConCin
Internally, we called it Project Deck.
The game itself is ConCin.
The project became a useful test for our design language because it forced a simple question:
Does this feel handled?
Not polished in the abstract. Not animated for decoration. Handled.
Cards should move like cards. The deck should feel like a deck. The interface should respect the memory people already have in their hands.
That is not easy to get right.
But it is exactly the kind of thing automation should help us explore. Agents can produce attempts. Human taste decides which attempts carry the feeling.
The lesson
Project Deck taught us a rule we keep using:
The metaphor is not a skin.
The metaphor is the spec.
If the product says "deck," the interface has to behave like a deck. If Prolifica uses a plant, the product has to propagate. If HARP is an instrument, orchestration has to feel playable.
The name, the motion, and the workflow need to agree.
That is the kind of software I want to build.
Software that feels like it has been made for hands, not only screens.
Happy building.