What I presented at Cursor Sydney
AIPRODUCT VIBINGMONEY MOJO
I recently spoke at the Cursor Sydney meetup at Atlassian, where I shared how I’ve been using Cursor and other AI tools to design and build side projects like Money Mojo and Product Vibing.
what I shared
1. Money Mojo - personal finance app
Money Mojo is my passion side project, a personal finance app I’ve been building.
It started from a real need. I wanted something more useful than a spreadsheet, with better visibility of spending, patterns, and where to make changes.
2. Product Vibing
Product Vibing is another project I built, and a big part of my learning curve with AI product building.
It’s a free 5-minute assessment designed to help you understand where AI tools can be used along the product development cycle, and how you can level up your knowledge and skills with AI.
A lot of what I learnt building Product Vibing shaped how I approached Money Mojo.
Web: URL: productvibing.com
Web: github.com/github/spec-kit
Workflow diagram: github.com/github/spec-kit/discussions/468
4. AI website builder for early design exploration
I also use AI website builder tools to speed up early exploration and concepting.
One tool I’ve used is Readdy.ai, especially for quickly generating UI directions before refining designs further.
Readdy.ai: readdy.ai
Referral link (1200 credits): https://readdy.ai/invite/AZWkQwog3v0u
what I’ve learnt so far
Three lessons have really shaped how I work with AI:
1. PPP = Piss Poor Planning Planning upfront saves time and pain later.
2. No process = rework A simple process creates better momentum.
3. AI isn’t a shortcut Stay in the loop. AI works better when you bring the thinking.
final thought
For me, the biggest mindset shift has been this:
3. Spec-driven development with SpecKit
One of the biggest shifts for me was moving to a more structured process before coding.
AI can build quickly, but it can also build the wrong thing quickly.
That’s why I’ve been exploring spec-driven development using SpecKit. I like the discipline of working through the steps clearly before implementation:
spec → plan → task → implement
The optional clarify, analyze, and checklist steps are also incredibly useful for catching gaps early.
Context first.
Code second.
AI is incredibly useful, but the quality of the output still depends on the quality of your inputs, decisions, and process.
Join the conversation
Visit my LinkedIn post: