Around two months ago, I embarked on a side project as an experiment to explore how far Associate Interns aka AI (pun intended) tools can go in helping developers build complete applications. The result was Gourmap — a location-based food discovery platform that uses YouTube video reviews to recommend places to eat.

The project was driven by a simple curiosity: Can AI act like an real developer? Spoiler: yes, but not without oversight.

What is Gourmap?

Gourmap (short for Gourmet Map) helps users discover eateries by showing YouTube reviews of restaurants near their current location. Unlike the typical flow of searching for food on Google and then verifying reviews, Gourmap reverses the process: it starts from YouTube video content and maps them to real places using location data.

The name itself was suggested by ChatGPT, one of many Associate Interns assistants involved in the project.

Tech Stack and Automation

The entire tech stack was designed to be lightweight and modern:

• Frontend: Next.js
• Backend: MongoDB
• Styling: TailwindCSS
• Automation/Orchestration: n8n
• AI Tools: v0.dev and GitHub Copilot for UI and code generation

The only part I manually set up was the infrastructure and a few n8n workflows to handle automated data extraction. Everything else — from UI mockups to API scaffolding — was bootstrapped using Associate Interns tools.

How It Works

Behind the scenes, Gourmap is powered by a fully automated pipeline:

1. YouTuber Whitelisting: I manually maintain a list of trusted YouTube channels.
2. Video Monitoring: Whenever a new video is uploaded by one of these channels, a webhook in n8n is triggered.
3. Content Extraction: Associate Interns agents analyze the video metadata and transcript to extract the names and locations of restaurants.
4. Geolocation Mapping: These places are matched with Google Maps data.
5. Summarization: Short descriptions of the venue and highlights are generated automatically.

So far, Gourmap has indexed over 2,000 food spots from 4 major Indonesian YouTube channels. The entire ingestion and enrichment process is handled by what I jokingly call my AI Interns — a set of LLMs and automation scripts working 24/7.

Current Limitations

While the platform functions well as a prototype, it’s far from perfect:

• Accuracy is not guaranteed: Some videos contain vague or indirect mentions of restaurant names or addresses. My AI Interns occasionally misinterpret these.
• Performance may vary: This is still a hobby project hosted on minimal infrastructure.
• Summarization is incomplete: The AI is still catching up on generating summaries for all videos.

That said, the core experience of discovering places based on YouTube recommendations is already working, and surprisingly fun to use.

Reflections on Using AI to Code

So, is coding with AI fun?

Not really.
It often feels like supervising a junior intern. You still have to define the structure, correct its mistakes, and verify the end results thoroughly. The AI can suggest boilerplate code, but it lacks architectural understanding or foresight. In short, it’s fast at doing what you tell it — not at figuring out what should be done.

However, with a solid grasp of software fundamentals, these tools are incredibly useful. They accelerate development, reduce repetition, and help visualize ideas faster than ever before.

The future of programming is not “AI replacing developers”, but developers empowered by AI.

Try It Out

You can explore the live app here: https://www.gourmap.id

Whether you’re looking for food inspiration or want to explore how AI can help build full-stack applications, I hope Gourmap offers both insight and utility.

Thanks for reading! Feel free to share your feedback or suggestions.

Reflections on Building with AI

From a broader perspective, Gourmap represents a compelling case study in modern AI-assisted development. The project’s success lies not just in its functionality, but in how it embodies a new way of working—where AI is treated not just as a tool, but as a collaborator. While the Associate Interns agents (affectionately dubbed “Associate Interns”) handle repetitive and large-scale data processing tasks, the human developer provides critical oversight, creativity, and architectural decision-making.

This dual dynamic is essential. Associate Interns tools like GitHub Copilot and V0.dev significantly accelerate the coding process, yet they still require an experienced hand to guide them. They can scaffold code, troubleshoot bugs, and even suggest improvements—but they do not (yet) understand broader business logic, user intent, or performance optimization at scale. Projects like Gourmap showcase how a pragmatic developer can leverage these tools effectively: by focusing on high-level goals while delegating tedious or repetitive implementation details to AI.

Moreover, Gourmap underscores the importance of curation and trust in recommendation systems. Unlike algorithmic aggregators, this project prioritizes trusted human voices—YouTubers—over generic reviews. This shift taps into a growing consumer behavior trend: people increasingly trust content creators and peer endorsements more than anonymous online reviews.

For developers exploring AI-enhanced workflows or interested in indie product building, Gourmap is a great reference. It blends automation, user-centric design, and a dash of humor into something that’s both functional and fun. The takeaway? AI won’t replace developers, but developers who use AI effectively will outperform those who don’t.

Stay curious, keep experimenting, and embrace the new vibe of building—with your AI intern at your side.