Struggling Forward: Cursor, Mock APIs, and the Weight of Self-Doubt

54 Hours of Code — and Still No Features Shipped

This week I clocked in 54 hours of coding, with 26 hours dedicated to a new project (I'll reveal more soon, stay tuned). And yet, despite the time and effort, I have nothing production-ready to show. I’m still chasing that elusive MVP.

I ended up rebuilding the core structure of the mock API—again. The previous version had become a maze of overcomplicated logic. Cursor helped me scaffold things faster, but to be honest, it often feels like managing a swarm of junior developers: fast output, but little understanding. What it gave me in speed, it lacked in clarity.

Now I’m deep in the refactor phase—rewriting, untangling, reshaping. It’s slower than I expected. And to be candid, I don’t know if I’m the only one feeling like this, but using AI tools has made me feel more like a follower than a builder. I'm second-guessing every line. Ironically, it's taking me just as long—if not longer—than doing it all manually.

The Battle With Myself

This week hit hard. I’ve questioned my worth as a developer multiple times. I even caught myself drafting a message to my boss: “I’m done—I’m just not good at this.” Between the mental fog of burnout and the physical toll of getting sick, I felt defeated.

But I didn’t quit.

I kept pushing forward. I dug into Cursor’s rule system, took apart the problem, and rebuilt pieces step by step. I’m not there yet, but I’m still in the fight.

If I could give my past self advice, it would be: "Keep going. Endure now. The payoff will come later."

Tools, Time, and Tiny Wins

• Cursor – Powerful, but not magical. It’s only as good as the prompts I give it, and prompt engineering is its own learning curve.
• Obsidian + WakaTime Plugin – Helped me document ideas and track the architectural decisions behind the scenes.
• Neovim – I spent 15 hours here, tweaking, configuring, and debugging. It remains my sanctuary for focused work.
• While listening to the DevRel podcast from Cursor, I realized something humbling: even juniors are expected to understand the product and business logic. That struck a nerve. Cursor can write code, but it can’t understand the problem space unless I do. That’s on me.
• Next steps? You’ll hear more soon about the new project I’m building in collaboration with ivoyant. We’re starting small: a tight MVP for a handful of users, then iterating with a Kaizen mindset—continuous, mindful improvement.

What Comes Next

• Ship the MVP, no matter what. It doesn’t need to be pretty. It needs to run.
• Get better at prompting. I’ll begin tracking what works and what doesn’t—turning experiments into experience.
• Rebuild confidence. I don’t need to be perfect. I need to show up.

WakaTime Snapshot

• Primary Language: TypeScript (20h)
• Editors of Choice: Neovim + Cursor
• Top Project: New application (coming summer 2025)
• Global Rank: #1336 on WakaTime

Goal: break into the Top 10 by next month.