top of page

Everyone’s Getting Better At Using AI — But What Does “Vibe Coding” Really Mean Now?


As we move through Q2 of 2025, one thing is clear: AI is no longer a novelty—it’s a toolkit. And across engineering teams, startups, and enterprise labs, something interesting is happening: everyone’s becoming a builder.

What used to require months of code is now possible with prompts and smart workflows. But this shift isn’t just about tools—it’s reshaping what it means to build, lead, and even code.

Engineers vs Non-Engineers: Is That Line Still Real?

There was a time when the difference between a builder and a manager was clear: one shipped code, the other made slides.

Today, that line is blurry.

We’re seeing non-technical founders shipping MVPs with GPT-powered workflows. Designers writing backend logic. Product managers building agents. The gap between knowing how to “engineer” and how to “deliver outcomes” is closing—and fast.

But let’s be real: engineering still matters.

Coding might be getting commoditized, but engineering—the creative structuring of systems, logic, and interfaces—remains deeply human (for now). As Shawn “Swyx” Wang put it in a recent podcast, “It has always been valuable for people to keep the pulse on what builders are building.”

And as the bar lowers for entry, the demand for excellence rises. That’s why communities of builders—people who still touch the keyboard—matter more than ever.

“The people who have hands on a keyboard still need a place to gather.” – Swyx

Let’s Talk Tools: Model Context Protocol (MCP)

Diagram showing MCP Host/Client (IDEs, Apps) and MCP Server. Arrows indicate "Tool Call" and "Tool Call Result or Context" between icons.

One of the standout ideas in that same conversation was the Model Context Protocol (MCP)—a framework for giving AI models structured, actionable context.

It’s open-source, hosted on GitHub, and aims to do one thing well: connect large language models (LLMs) to real context via client-server architecture, integrations, and structured environments.

Think of MCP as the infrastructure layer behind smart agents and autonomous workflows.

Here’s what it enables:

  • Model flexibility (you’re not tied to one model)

  • Better communication between tools and models

  • Smarter context injection

It’s early days for MCP, but it points to something deeper: context is king in this new AI world. Without context, even the best model is guessing.

The Rise of AI Agents – and the “IMPACT” Framework

A woman uses a laptop at a desk, interacting with a robot. Surrounding icons suggest technology and AI. The background is light blue.

AI agents aren’t just theory anymore—they’re starting to become real, practical tools in the developer workflow.

Swyx summarized agent architecture using an acronym: IMPACT

  • I – Intent + Intensity

  • M – Memory

  • P – Planning

  • A – Authority (domain-specific expertise)

  • C – Control flow

  • T – Tool use

An agent isn’t useful because it “talks”—it’s useful because it does. And that only happens when you combine LLMs with structured memory, intentional planning, and real-world tools.

That’s why Cysparks is increasingly focused on agent engineering through Aivara as part of our upcoming services.



So... What Is “Vibe Coding”?

Man gesturing on a green and white background. Quote: "There's a new kind of coding I call 'vibe coding'..." - Andrej Karapathy. Education Next logo.
credits:EducationNext

There are two schools of thought here:

  1. Vibe coding as expertise – You understand what code should look like and use AI to accelerate or scaffold it.

  2. Vibe coding as vibes – You prompt your way through a build with no technical foundation and hope AI gets it right.

The first is leverage.The second is chaos.

At Cysparks, we believe “vibe coding” is not a shortcut to skip learning—it’s a fast lane for those already on the road. AI helps you build faster, not blindly.

What This Means for You

Whether you’re a startup founder, CTO, designer, or a student figuring out where to start—this shift affects you. AI is not coming for your job; it’s coming with your next job.

You’ll need:

  • Adaptability – to learn new tools fast

  • Engineering literacy – even if you’re not a developer

  • Clarity – on what outcomes you’re driving

The future belongs to those who can bridge vision and execution.

And as we build from Kenya to the world, this matters even more. We’re not just consuming AI—we’re shaping it, remixing it, and applying it to real challenges.

As Harvard Business Review puts it:

“The need for professionals to adapt and continue to learn and grow becomes more imperative.”

Let’s keep building. Smartly. Intentionally. And maybe, just a little vibey too.




Written by The Techonomist, & Samuel Mogul.

See you in the next one.



Comments

Rated 0 out of 5 stars.
No ratings yet

Commenting on this post isn't available anymore. Contact the site owner for more info.
1/2

Be the first to know!

Thanks for subscribing!

1/3
How useful was this article to you?
Don’t love itNot greatSatisfiedReally goodLove it
Found a mistake in this article?
Report it to us
What is the issue about?

Thank you for submitting your issue. We review it and get back to you shortly. If you any further questions or concerns, feel free to reach out.

bottom of page