Your Life

Inspiration

I am a 64-year-old retired pilot with no software or coding experience. Aviation taught me that important decisions must be supported by evidence, exceptions must remain visible, and a human must retain final responsibility.

In everyday life, however, information is scattered across bank statements, invoices, messages, folders and people’s memories. I wanted a clearer and more accountable way to turn that evidence into action.

What it does

Your Life is an auditable AI control-room prototype for personal and small-business operations. It brings cash, commitments, evidence and continuity into one view while keeping Personal, Your Biz, KSJ and Legal records in separate silos.

The system links operational events to evidence, identifies missing or conflicting records, assigns exceptions to human owners and produces concise decision briefs.

An AI decision brief cannot be treated as controlled while unresolved evidence exceptions remain. The interface clearly explains what evidence is missing and what a human must do next. All information shown in this prototype is labelled demonstration data.

How I built it

I described the operational problem and control principles in plain language. Codex and GPT-5.6 helped me turn those ideas into product architecture, interface designs, evidence-control logic, tests and a deployable responsive web application.

I reviewed the results, chose the design direction, tested the application on my iPhone, requested corrections and retained final authority over every decision.

The demonstration uses deterministic sample data rather than real financial records or live bank connections.

Challenges

My biggest challenge was learning an entirely unfamiliar technical process: GitHub, repositories, testing, deployment and mobile demonstration recording. I also had to make sure the product never presented an AI-generated conclusion as controlled when supporting evidence was incomplete.

Accomplishments

Without previous coding experience, I created and publicly deployed a working, tested prototype. Its central achievement is not merely producing answers—it preserves evidence boundaries, exposes exceptions and keeps humans accountable.

What I learned

I learned that a non-programmer can use Codex as an engineering collaborator by stating the desired outcome clearly, reviewing each result and insisting on verification.

What’s next

The next phase would add authenticated evidence ingestion, encrypted storage, role-based approvals, append-only corrections and AI outputs constrained to cited evidence objects.

Built With

• codex
• github
• gpt5.6
• next.js
• react
• typescript