My Take on 6 Weeks in a Vibecoding Accelerator: Week 2
Eight weeks ago, I was deep in Week 2 of the Lovable Shipped Accelerator - a 6-week sprint to vibe-code an app, ship it, and (in theory) walk away with $3 million worth of tools, expert guidance, and maybe a little investor attention.
If Week 1 was chaotic excitement, Week 2 was all about pivots, reality checks, and the first big “why did I do that” lesson.
From chatbot to clean slate
Last week, my fixation was onboarding. I had this vision of a sleek, full-screen experience that would set the tone for my app.
What I got instead was a chatbot.
Not an AI-powered, magical, personalized chatbot — just a logic-based question flow that looked nothing like I imagined.
I kept thinking, If I just keep refining this, it’ll get there. So I poured time into it, rearranging flows, cleaning text, tweaking styles.
Then Felix Haas - our Week 2 design guest - gave a session on iterative improvement, user feedback, and focusing on what actually drives the product forward. He left us with prompt guides, a “prompt hack” sheet, and a buzzword guide. More importantly, he left me with a shift in perspective:
“If the foundation isn’t there, no amount of polish will make it work.”
It hit me: my onboarding couldn’t actually onboard anyone because there wasn’t enough built to point them to.
I scrapped it and started fresh.
The first big lesson
Looking back, this was my first true accelerator pivot - and one I wish I’d made sooner.
Lesson learned: Don’t waste valuable build time on features that depend on other features you haven’t built yet.
Instead, I redirected to:
Creating a simple, clean landing page for logged-out users
Mapping out the onboarding overview without worrying about branding yet
Ensuring the main app flow worked end-to-end before adding any “extras”
The perks reality check
Turning in Week 1 homework unlocked the promised “$3M worth of perks.”
Reality?
A few fantastic full-year licenses and genuinely valuable tools
Plenty of discounts - some only useful if you upgraded to a higher tier
A few offers identical to what was already available on the companies’ websites
Not the $3M treasure chest I pictured, but still several gems I’ll actually use - now and in the future.
Progress under the hood
Even with the onboarding detour, I still made solid progress this week:
Added several new Supabase tables in preparation for the main data setup (still in construction)
Integrated a waitlist directly into the app, replacing my temporary one on another site
Tested all homepage functions and button flows for a smoother user experience
Built a more readable onboarding workflow (while still wrestling with authentication issues - Lovable and Supabase were not playing nice)
The unexpected win
One of the best parts of the program is feedback day. Every week, we’re required to give detailed feedback to other participants. In return, we get their feedback on our own work.
The encouragement was great - but the real takeaway was that people were genuinely interested in my product. That gave me the push I needed to push through the bugs and keep building.
Looking back:
If Week 1 was about excitement, Week 2 was about course correction. I learned that “building in public” isn’t just about showing wins - it’s about showing the pivots, the scrapped ideas, and the moments you realize your priorities were upside down. And while I’m building in public on delay - it’s still public.
If I had to do it again? I’d still make the pivot on onboarding...Just sooner.
If you want to see how Invoiceabill has evolved since those early weeks, you can check it out here: invoiceabill.com
