Founder, YMDB. Growth strategist at @JhavtechStudios. Helping startups & businesses fix & grow their apps. ex-TWTR, MSFT, YHOO. DM me!
Melbourne, Australia
Joined May 2007
- Tweets2.5K
- Following957
- Followers2.4K
- Likes446
Rick Mulia retweeted
Cal AI would have scaled to $200M ARR if I had access to this 2 years ago 😭
FunnelFox just shipped their MCP, and it's the #1 tool I'd want if I were running an app today
(especially if I’m trying to avoid apple’s 30% fees)
Just connect it to your AI assistant and build or tweak any web funnel with whatever you’ve got:
> a funnel URL you like as inspo or a Figma design you like
> your own funnels that already convert
> exactly what competitors are running (pulled straight from their Market Radar)
Everything it builds is a real FunnelFox funnel, with the quiz, paywall, checkout, and web2app logic already wired in
Now you just type one line and you watch Claude or any other AI update or create your funnels in minutes:
"build a funnel like this one, but for a male audience"
"adapt my top funnel for all my geos"
"change prices across every checkout"
Our edge at Cal AI was doing the simple things faster than everyone else
This is how you win in consumer apps in 2026 and beyond
See how it works for yourself: funnelfox.com/?utm_source=ja…
🤝 Paid partnership
Product Page Headers & new Search Results creative assets are easy, free opportunities for any app owner to increase their App Store presence.
ALT These are the major changes to App Store presence that can impact your share of downloads
When taking over a vibecoded project the gut instinct may be to start from scratch, especially with a design prototype.
Where possible, build on top of what the customer has put their blood, sweat & tears into. You'll get greater buy-in when you want to evolve the product further.
With the launch of the iPhone Duo there's new free opportunities in the App Store to promote you app.
Product page headers and larger search results presence gives you more surface area to highlight your app across all App Store fronts.
No iPhone Duo purchase required.
Building a vibe-coded app changes what founders expect from dev partners.
It's no longer a traditional client-dev relationship.
Co-building allows founders to keep their freedom & ability to iterate. A full-stack dev team is there to help launch or get to the next milestone.
Building & launching an app means setting up dozens of things you'll barely think about again, until you need them.
API keys, secrets, access tokens, passwords. Vibe coding tools store some of it, but good documentation and processes become essential as your app and team grow.
It's ok to build an app that's just for you, your family or a small circle of people. The metrics of building an app these days make this realistic. You can do away with marketing costs, feedback loops & iterate just for you. It doesn't have to be profitable to meet your needs.
Custom Product Pages in App Store are an undervalued growth play. Run targeted Apple Ads highlighting different user benefits. A/B test these until you find the ones that work for your app. Can be added without a new app update.
Devs will argue that vibe coding doesn't produce scalable code.
Agreed. And that's not the point.
A non-technical founder can now get further with their idea.
What they need now are experienced humans who can give them advice on setting themselves up for future scalability.
The best way to avoid AI slop when building an app?
Prompt like a designer & developer. The language that coders & designers use allows specificity & limits misunderstanding. Mistakes will still happen but describing an issue clearly will help you get to shipping product faster.
Don't product plan based on what comes up in your feed.
You see a cool app feature as you doomscroll and you think this will be a game changer for your own app.
What works for one product and one audience may not work or be so important for another audience.
What's worse than feature creep with vibe coding?
Time creep.
With unlimited dev you build all your ideal features. The problem is you have to spec, test, & launch each of these features.
You become an AI zombie, testing & fixing your agent's product. This is a huge time suck.
I moved my biggest app project from Lovable to Claude. Needed to get past the endless loops of errors, fixes & testing. Also had to refine the design beyond Lovable's standard templates.
I miss the realtime preview window but the benefits of moving to Claude is a bigger plus.
Everyone can build an app now these days.
Most people who build one won't turn it into a business. They'll have a working demo and no idea what happens after "it works."
If you've got a vibe-coded prototype, ask yourself which problem you're actually solving next.
Vibe coding gets you to a prototype fast. It doesn't get you out of Terminal or through the app stores. Shipping still means understanding your codebase, fixing build issues, and navigating Apple and Google review. That's where products become apps.
It's easy to get stuck into an infinite loop of QA with vibe coding.
You keep testing, with AI saying they've fixed the error and asking you to do "one last push".
Shorten these loops by highlighting all the differences you notice in various tests, no matter how small.
How long from prompt to a working prototype, by tool:
- Bolt: 28 mins
- Lovable: 35 mins
- Replit: 45 mins
Each has their pros. Lovable gives you portable code. Replit is best for proper deployment. Choose the tool for where your product is headed, not just how fast it is.
How long from prompt to a working prototype, by tool:
- Bolt: 28 minutes
- Lovable: 35 minutes
- Replit: 45 minutes
Each has their Pros. Bolt is fastest to a demo. Lovable gives you portable code. Replit is built for long term development. Choose the tool for where your product is headed, not just how fast you can build it.
An App Store rejection can be a kick in the guts.
Getting that email is deflating.
Don't let it get to you.
In most cases the rejection reasons are basic technical or process issues. Sprint to fix those problems & don't stuff in more features. Ship then iterate.