🗞️ Issue 65 // ⏱️ Read Time: 6 min

Hello 👋

It all started with a tweet (well, a post on X, formerly Twitter) on February 2, 2025: “There's a new kind of coding I call ‘vibe coding’, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists.”

In other words, the expression Vibe Coding is about 2 months old. So, what does it mean? Sit back and relax. Once you’ve finished reading this newsletter (in 6 minutes, about the time it takes for you to take some good sips on your beverage of choice), you will know exactly what’s going on.

To our faithful newsletter readers

Before we continue to talk about the future of AI and vibe coding: We would love to meet you IRL in Lisbon at the end of April! See all the information below.

The Rise of Vibe Coding

So, imagine having a vision of an app or software you want to build and vibing your way into it with the help of an AI tool. No coding needed. This space is developing so fast that you wouldn’t even be able to keep up if your name were Usain Bolt.

Can it really be that easy? What about all those years in university, and the blood, sweat, and tears it takes for developers to build great products? Well, let’s take a closer look.

The Vibe Coding Tools: It’s so easy

Some quite recent products in the field of AI are changing the way we imagine, build, and grow products (and organisations!) moving forward.

You know how cameras are now an integral part of most smartphones? We got here from darkrooms via digital cameras. Software development is part of the same “democratization wave”. Traditional coding was expert-driven and hard to understand. Around 2010, we reached the low-code/no-code area, with tools like Zapier emerging on the market. We reached AI-assisted development with tools like Github Co-Pilot, and landed where we are today: The Vibe Coding era.

Platforms like Windsurf, Cursor, LoveableDev, and Replit promise that anyone can take their app or software idea to execution within minutes. And yes, to a certain extent, that is true: Team Lumiera recently tested this ourselves and within an hour, we had tools that would have taken us weeks to build. It is easy. Is it…too easy?

The Not So Good Vibes

Think about it: Entrepreneurs are now empowered to validate ideas, without hiring expensive developer teams. Without learning Python, we can solve problems within our domain. Those who, for some reason, got the idea that engineering is not for them can now create their own solutions. All of this is, of course, great news.

So, are there any downsides to this? Yes. Definitely. And in true Lumiera fashion, we look at the good and the bad. Analysis and context. That’s how you become AI literate: Understanding the nuances of topics like these.

When not understanding the system becomes a problem

By vibe coding, we’re essentially removing the connection between creator and medium. In this case, it’s between developer and code. You can compare that to a chef who doesn’t taste their ingredients, or a sculptor who doesn’t touch clay. Would you trust a hairdresser who doesn’t know how to use a pair of scissors?

Essentially, it’s a matter of understanding systems on a more fundamental level. Because, just like in life, things in tech break. Having a more solid understanding of systems means that solutions are more elegant and maintainable, and stand the test of time.

Blind trust is the death of reasoning

Humans are lazy. Yes, you are too, and so am I. This means we are inherently biased to trust that AI will be right, especially when you’ve built an app where every single step towards it would take you hours to understand. When the code cannot be understood, how can we mitigate the risk of AI hallucinations? How can you verify that you get the outcome that you are actually trying to achieve?

Tech debt

We believe that the concept of technical debt will soon become part of the standard vocabulary of most executives. Because they will take the consequences of speedy delivery of apps, products, or other tech solutions, where the speed of execution is prioritised over the speed of quality. In the short term, vibe coding can be used to deliver a solution fast, but it may increase future costs and complexity if left unresolved.

So what, is pressing PAUSE a good idea?

After speaking with a dear friend of ours about the problems with vibe coding, he asked a fair question: “So what, should we just not use these tools now?” To be fair, we’re just getting started. Embrace the change or become irrelevant. Our approach is that rather than just quitting vibe coding altogether, we can approach it responsibly. Like that, we get the vibe coding benefits, but can mitigate the main risks mentioned above. These are some main things that will preserve the craft while embracing the innovation.

Vibe coding tips for individual creators

1. Learn the fundamentals first. Build a strong foundation in programming concepts before relying heavily on AI. Don’t hire engineers if they cannot code without the support of an LLM.

2. Use AI as a collaborator, not a replacement. This one is a classic, because it’s true.

3. Understand what the AI produces. Take time to read and comprehend generated code before implementing it.

4. Challenge AI outputs. Don’t accept the first solution. Ask, “Is there a better way?”. “What are the factors you’ve considered when proposing this solution?”

5. T-shaped skills: Be a generalist and a specialist. Develop deep knowledge in one area, and a broad understanding across many.

Vibe coding tips for teams and organizations

1. Establish robust review processes for AI-generated code. Don’t skip quality assurance just because AI wrote it. These types of processes are becoming more, not less, important.

2. Create balanced teams: Mix AI enthusiasts with different skillsets with traditional developers. Checks and balances

3. Learn system thinking: Invest in education that emphasizes system thinking and architecture, not just prompt engineering and vibe coding.

4. Diligent documentation: With less human-written code, thorough documentation becomes even more crucial.

5. Focus on the problem you are solving: Restructure around problem spaces, not technical specializations.

Wrapping up

If you made it this far, thanks for reading (or listening) all the way down here. Our take on this topic is that the rise of vibe coding doesn't signal the end of traditional software development but rather a transformation in who creates technology and how.

We would love to hear your take on this, too. Have you heard about the concept of Vibe Coding before? Have you tried any of the platforms promising that anyone can take their app or software idea to execution within minutes? Are you planning to change your ways of working, or maybe your ways of hiring? Let us know - we appreciate every interaction that we have with you through this newsletter.

Until next time.
On behalf of Team Lumiera

Emma - Business Strategist
Allegra - Data Specialist

Lumiera has gathered the brightest people from the technology and policy sectors to give you top-quality advice so you can navigate the new AI Era.

Follow the carefully curated Lumiera podcast playlist to stay informed and challenged on all things AI.