From Vibe Coding to Real Impact: What We Learned at Our First Cursor Meetup
Connect with me on LinkedIn: jpmarindiaz
Last Tuesday, 40 developers, designers, and curious minds packed into our Datasketch space in Bogotá for something special: our first Cursor meetup. What started as a simple idea to share experiences with AI-powered coding turned into something much more revealing about how we’re all adapting to this new era of development.
The Energy Was Immediate
Walking into the room, you could feel the excitement. We had everyone from seasoned backend developers to designers who typically run away from code, all united by one thing: curiosity about how AI is changing the way we build things.
The diversity was striking—DevOps engineers talking to product designers, data scientists sharing tips with frontend developers, and people who had never programmed before sitting next to those with decades of experience. What became clear quickly was that Cursor isn’t just changing how we code; it’s democratizing who can code.
What “Vibe Coding” Actually Means
We spent time defining this term that’s been floating around the community. One participant perfectly captured it: “It’s like having a friend who’s a good programmer sitting next to you, and you’re just telling them ‘do this, no I don’t like that, now change this.’”
But here’s what surprised me: vibe coding isn’t just about letting AI do everything. The most successful practitioners in the room were those who understood that they were becoming managers of AI agents—breaking down problems into smaller, manageable pieces and orchestrating solutions rather than just accepting whatever the AI spits out.
Francisco, who’s building training courses for accountants, put it best: “I changed from thinking about what function to write to thinking about what context I need to give the AI to solve the problem correctly.”
The Real Power: Modular Management
One of the most interesting insights came from watching how experienced developers were adapting. They weren’t just using Cursor to write code faster—they were fundamentally restructuring how they approach projects.
I demonstrated how I work on complex projects by having multiple Cursor windows open, with each one handling a different module. One window exports functions, another consumes them, and I act as the mediator, asking Window A to check what Window B is doing and ensuring consistency across the entire system.
This approach revealed something important: we’re not just coding anymore, we’re orchestrating. And that requires a different set of skills—ones that emphasize architecture, planning, and clear communication over pure technical implementation.
When AI Meets Real Projects
During the session, I worked live on a real Datasketch project—our biodiversity information system for the Humboldt Institute. When deployment errors came up (the kind that used to make me avoid touching JavaScript projects), I simply pasted the error into Cursor and watched it diagnose and fix hydration issues in our Next.js application.
What made this powerful wasn’t just the speed—it was that I could finally contribute meaningfully to parts of our codebase that were previously off-limits to me. Suddenly, the barriers between “my expertise” and “their expertise” started dissolving.
The Honest Challenges
Not everything was roses and rapid prototyping. Several participants shared cautionary tales:
- One founder had to shut down his application because the code became so sprawling and unstructured that even Cursor couldn’t understand it anymore
- Security vulnerabilities were mentioned as a real concern when you don’t fully understand what the AI is creating
- The temptation to accept everything without review led to projects that worked but couldn’t be maintained
The consensus was clear: vibe coding works brilliantly for exploration and prototyping, but production applications still need human oversight and architectural thinking.
Beyond the Hype: Practical Insights
Some of the most valuable insights came from the practical tips shared:
From Henry (Backend Developer): “I use the @ symbol to link specific files when working on large codebases. It keeps Cursor focused on relevant context instead of changing random files.”
From Nicolò (API Integration Specialist): “I ask Cursor to write tests that are designed to fail, then have it learn from those failures and modify the code. It’s like teaching it through iteration.”
From our DevOps expert: “I haven’t written a YAML file in months. I just give Cursor documentation and let it handle infrastructure as code. What used to take weeks now takes hours.”
What This Means for Small Teams
For organizations like ours at Datasketch—where we’re building data democratization tools for small teams—these insights are particularly valuable. We’re not just making our own development faster; we’re seeing how our future clients might interact with data and technology.
During the meetup, we live-coded a chatbot logging system for our data visualization platform. In just the time we were talking, we had a working API that could track user interactions with our AI-powered data tools. This wasn’t just a demo—it was real infrastructure we’ll use to improve our services.
The Human Element Remains Central
Perhaps the most important realization from the evening: despite all this AI capability, the human element became more important, not less. The developers having the most success weren’t the ones just prompting and accepting—they were the ones who had become better managers, better communicators, and better system architects.
As one participant noted: “You still need to understand the fundamentals. Cursor makes you more powerful, but only if you know what you’re asking for.”
Looking Forward
What excited me most wasn’t the technical capabilities we demonstrated—it was the democratization we witnessed. Designers who previously avoided code were exploring APIs. Non-technical founders were building prototypes. Experienced developers were tackling infrastructure they’d never touched before.
This isn’t just about productivity gains (though those are real). It’s about expanding the circle of who gets to build things, who gets to solve problems with technology, and who gets to participate in creating the tools that shape our world.
Join the Conversation
We’re planning more gatherings like this—some focused on Cursor specifically, others exploring AI tools for non-programmers. If you’re in Bogotá and interested in how AI is changing creative work, data analysis, or just curious about what’s possible when technical barriers start dissolving, we’d love to have you join us.
The future of development isn’t about humans versus AI—it’s about humans and AI learning to work together. And based on what we saw Tuesday night, that future is going to be pretty exciting.
Want to try the kind of data tools we’re building at Datasketch? Create and account and check out how our AI-powered visualization platform can turn your spreadsheets into insights without any programming required. Because whether you’re vibe coding or just trying to understand your data better, the goal is the same: making complex things accessible to everyone.
