To start learning about vibe coding I started with building a triathlon calculator as it was something I knew and used them myself. It has grown, iterated and lots of learning. I’m the primary user and once I’m getting value from it I will speak to others to validate assumptions.
My current flow:
- Use https://www.runningcalendar.ie/ to find my next races
- Lookup a race distance
- Lookup races by dates
- Before racing I would use a calculator or set pace goals
- Use Triathlon Calculator
- Set a time goal in notion
- Save results of a race in a notion table with notes
- Take screenshot of table and paste on results page
- Create blog posts as race reviews
- Read peoples FB, Reddit posts
- Take inspo for future races
- Add future race to notion table
The value add was to create a community around this and then look at monetizing later on based on needs. I never did some proper competitor research on what is out there for this so discovering tools as I progress.
Started as a triathlon calculator called TiPacing
Vibe coding this with lovable was so much harder than you might think. It constantly changed the wrong items. A new model came out and it got so much easier.
Biggest problems
- Once I added a second calculator it got 10X harder to build
- AI Coding tools were more single page / use at the time
Then a race calendar where you select the race that you are setting a time prediction called NegativeSplits.ai
First thing was building out the AI Triathlon Pace Calculator – Save Predictions into negative splits and embedding on this site to drive traffic. Idea was could reach out to tri clubs and offer a free Vibe Coding Website Free – Prompts for non-Techies and include the calculator which would add a backlink.
I spent a lot of time on data integrity so others could save race result against a instance in the past. Also, didn’t want database full of poor race details so tried import all Ironman, Challenger and local Perth races so people could test.
Biggest problems:
- I don’t discover races from a map
- I don’t want spend time looking up sites to find races
- Race Data integrity with archive of instances
- Database of all races is extensive
I have discovered numerous possible same tools some new vibe coded and others I just never knew existed. In reality I have only ever searched for races not a app to save my races. The barrier to entry to create directory websites has been heavily reduced.
Now I’m thinking actually what I want is a Personalized Email
I don’t want to go rooting through a calendar to find next race. The app was never supposed to be just a fancy a calendar with a map. The idea of a newsletter personalized to my interests sounds interesting and plausible.
- Select distances I’m interested / location
- Send me what’s coming up
Was negativesplits a massive example of scope creep. Do I just want to upgrade my results page on this website and add a race bucketlist feature!?
I got to learn
- Data scraping tools like https://www.firecrawl.dev/
- Saw first hand now whats possible with latest coding models (Sonnet 5, GPT5-6)
- Remembered the consequences of tech debt
- Start with use cases, personas and data models.
- Problem you keep adding items like clay instead of sculpting something.
- Cost me 2 domains ($40) and 5 months on Base44
If I was to build a version not worrying about bad data, event management, finding races it could look a lot different as takes out a lot of complexity. Even bringing back the original negative splits homepage graphic
Probably main conclusion I love using Base44 and committed to a yearly plan with them. Maybe next big project is migrating shanedrumm.com to base44.