Episode 04-04: Vibe Coding – The Bad

Posted on by

Welcome back to Wrenches, Gadgets & Hacks! This is the fourth episode of Series 4, The Great Unknown, and we’re riding the AI wave known as vibe coding. But as this episode proves, the vibes aren’t always so rosy.

After exploring the good side of vibe coding, Carl and John dive into the messy middle, “The Bad.” This episode takes on the pitfalls that surface once the AI excitement fades and the real coding challenges begin.

They share stories from the trenches: apps stuck in loops, burned-through API credits, context limits that wipe out progress, and the harsh truth about tokens, budgets, and overly confident “helpful” code. From Gemini to Cursor, ChatGPT to Claude, the hosts break down what happens when the vibe breaks, and what you can actually learn from it.

 And don’t forget to check out the companion blog on TheBIMsider.com: Vibe Coding: The Bad.

Be sure to stick around to the end, and learn about these points👇(just click the Click to Continue Reading and See Show Notes)

Vibe Coding – The Bad – Image by ChatGPT

Key Discussion Points:

When the Vibe Breaks:
Carl and John pick up where “The Good” left off, diving into what happens when the AI magic wears thin and the roadblocks hit. The episode opens with the realities of vibe coding once speed meets complexity.

Hitting the Limits:
John talks about how long chat sessions and bloated prompts caused Gemini to stop rendering his app altogether. The discussion highlights the frustrations of context limits and why some projects stall before reaching the finish line.

The Cost of Curiosity:
They explore how “cheap” experiments can turn expensive fast. From burned API credits to surprise token bills, the hosts break down how AI tools quietly rack up costs and why understanding token usage is crucial.

Version Control or Bust:
Carl and John stress the importance of backups, GitHub repos, and version snapshots after learning firsthand how quickly code can vanish or break when AI goes off script.

When AI Gets Too Confident:
The hosts share stories of Copilot and Cursor making bold “fixes” that broke entire projects. The takeaway: AI can be a great assistant, but it still needs supervision.

Privacy and Security Wake-Up Calls:
John explains how building his note app raised serious data privacy questions. He added encryption just to make sure no one—including himself—could read someone else’s notes.

The 80–15–5 Rule of Pain:
Carl introduces his “80–15–5” rule for vibe coding: the first 80% feels easy, the next 15% is a grind, and the final 5% can feel impossible. Finishing is where most projects fall apart.

Breaking, Learning, and Moving Forward:
Despite the “bad,” both hosts agree that breaking things is part of the fun. Each setback teaches something new about AI, coding, and patience—and that’s what keeps the vibe alive.

You can listen to the podcast, HERE, the Spotify player at the start of this post, or on your favorite podcast app. If you prefer visual content, be sure to check out the video on YouTube.

Note: We are aware of some audio issues in this episode, but we felt the content was too good (especially for an episode ironically titled “Vibe Coding – The Bad”) not to share. We apologize for the inconvenience.

Show Notes:

🧠 AI & LLM Tools

💻 Coding & Development Tools

⚙️ Development Practices & Concepts

  • Tokens – A key topic; learning how they work and how they impact costs.
  • Version Control – Emphasized as a safeguard against lost or broken code.
  • Encryption – Added by John to secure his note-taking app’s user data.
  • Authentication – Discussed in the context of Firebase and data privacy.

🫣 Sneak Peak

John’s Vibe Coded notes tool “Rollo Notes” from within Gemini

 NOTE: Turns out, AI doesn’t mind a little “bad.” It lent a virtual hand with this write-up. 💻 🤖

You can find all of the WGH links in one place with our 👉 BIO LINK

Also a BIG Thanks to Alex Grohl for our amazing intro and outro music “Indie Hey Song (Lost Fragments)” You can find more great royalty-free music from Alex on Pixabay

If you enjoy the podcast content and wish to show your support, consider buying the show a coffee ☕👉 Buy the Show a Coffee

Leave a comment