60% plus users are non coders

Perspectives about product interviews requiring to do prototypes during the interview process

Moats in an Ai age

1. Speed as strategy, make everything faster, learn, ship, iterate

2. Brand as a Most. Ai created stuff is generic. Brand design can capture emotions.

Self thought. I wonder if loveable help create the architecture as a deterministic input. Frameworks of architecture and then translating what the user typed in into the architecture sections.

He gives an example of how to create a prototype. He has found success in incremental building. Small units at a time. Similar to how we did our slack campaign result tool.

He makes a cost efficiency point. Being very specific ensures that the AI goes and modified only what it requires , reducing the tokens and reducing the cost.

His suggestion to Vibe Code, think in terms of 3 buckets:

Ui, Logic, Data

He is experiments with Spec driven development.

Specify, Plan, Tasks, Implement, Validation 

Self thought. To differentiate, should we build a Brand Guide, a mood board, style guide and Brand identity before we start building.

Self thought. To further differentiate, should there be a Business Context or a Business Brief. 

How about the Interaction style guide. Imagine if the interactivity had an emotion and had brand feel.

Self thought. The ai generic design allows you to design for the masses. And avoid custom designs that users might not know and get confused about.

Quick formula to rapid prototype 

Stitch.google.com, gives a platform to build the ui ux. 

Google ai studio to convert screens into real prototype.

After initial prototype, when interactions get complicated, Claude Code.

I wonder how does one debug a logic issue and find out where it went wrong? Because you have no clue about the code itself. Could we ask the code what are the possible locations it could have gone wrong? Is there some tooling which