Hi there, great to have you here!
Check out what is in this issue:
- Signal: The hottest new programming language…
- System: How to Create Packaging Images with AI
- News: Vibe coding is getting crowded
Signal
AI is rapidly taking over the world at a pace that is making everyone anxious. A very nice take on this was shared by my friend Guilherme Lima in his newsletter, where he explains we are still in the information era of this new AI innovation cycle. There is so much happening at the same time that it’s impossible to be clear on how it will all play out.
I have a take on this.
Even though I agree it’s hard to predict how things will work in the AI future, there’s one thing I believe everyone—regardless of their field—should be investing time in right now: prompt engineering.
It’s the foundation. And it’s very different from Googling something (even though they kind of look similar at first glance). Prompting is not about searching. It’s about shaping the machine’s behavior with your words. You’re not asking for a page — you’re giving instructions. It’s like casting spells in a new digital language. And that language… is plain English.
As Andrej Karpathy put it: “The hottest new programming language is English”
Let that sink in.
For decades, we had to bend to the rules of machines. C++, Python, Java — these were our bridges to talk to computers. Cold, structured, logic-bound languages. Now? The tide is turning. Machines are finally meeting us on our side of the bridge. They’re learning our language — well, me are making them learn.
This shift is enormous. It democratizes creation.
Anyone who can think clearly and write it down can build, automate, research, or teach with AI. No code required. Just clarity, creativity, and a little practice with prompts.
So yes, the future is fuzzy. But this one skill—prompting—is as close as it gets to a sure bet. It’s the new literacy, the interface between our brains and machines. The earlier you get good at it, the more powerful you’ll be when this next era solidifies.
Start now. Write better prompts. Understand how AI thinks. Practice the craft. Because when the dust settles, this might just be the most valuable skill you bring to the table.
System
Following this idea of getting better at prompting, this past week we (at MindMiners) were trying to create a generic packaging image for a research project we were running for a client.
Between many tries, we finally reached the generic image with pills that we wanted, but it left me with the impression that we needed a guidebook.
With that in mind, I wrote a full article to work as a guide to prompting to generate packaging images in ChatGPT (4o model): How to Create Packaging Images with AI
There are still things to improve, but the final results can be awesome.
More importantly, this process taught me something fundamental: prompting isn’t just about typing something random and hoping for magic. A good prompt is a well-structured brief.
You need to know what you want, describe it clearly, and guide the AI step by step—just like you’d brief a designer or developer. The clearer your intent, the better the output.
This is why prompting matters so much. It’s not just about AI — it’s about thinking clearly and translating that thinking into action.
In the article, I break down a basic structure that makes prompting for images way more effective. It works like a creative brief—and has four core parts:
- Scenario – Where the product is placed (e.g., a bathroom counter, a white studio background).
- Object & Material – What the packaging is and what it’s made of (e.g., a frosted glass jar, a cardboard box).
- Design Style – The visual vibe (e.g., minimalist and modern, vintage and elegant).
- Logo & Brand – The identity you want on the product (e.g., centered logo, no extra text).
You can also add optional extras like dimensions, angle of the shot, or lighting references if you want to get really specific.
Once you think this way, prompting becomes way less “random wish” and more “creative direction.” The AI responds much better when you feed it structured intent.
Check the full breakdown and image examples here 👉 How to Create Packaging Images with AI
Latest Highlights
Curated News
Vibe coding is getting crowded
This past week Google and Canva moved into “vibe coding”.
WTF is vibe coding:
Vibe coding is a programming approach where AI generates software from natural language prompts instead of manual coding. It shifts the developer’s role from writing code to guiding the AI with clear instructions and refining the output.
Core principles of Vibe Coding:
- Iteration: Output usually needs tweaking, prompting a cycle of testing and refinement.
- AI-generated code: Uses LLMs to turn natural language into executable code.
- New developer role: Focus moves to prompting, testing, and refining AI output.
- Accessibility: Lowers the barrier to entry, enabling non-programmers to build software.
- Prompting: Clear, concise descriptions guide the AI.
In one hand, Canva made a huge update to its platform, moving the traditional design first to AI first in design. Six main features where launched or packed together (because many things were already available), enabling users to generate interactive applications through natural language prompts.
Now you can create elements like calculators, quizzes, and mini-games without traditional coding, streamlining the development process for users across various skill levels. That is the feature related to “vibe coding”
Google Firebase, in the other hand, focused on another technical level of vibe coding, and therefore target customer – developers. Now developers can prototype and launch full-stack AI apps using only prompts and smart defaults — with Gemini guiding the process like an embedded co-founder.
Have I tried them? Yes.
To something useful? No.
I asked to create a mini app that receives a name as input a creates a meme with it. On Google, many inputs and corrections were needed to reach nowhere, while Canva made a good outcome with just the initial prompt.
Here you can try NameMeme at Canva.
Here is my take on Vibe Coding
There’s a growing hype around vibe coding — the idea that anyone can build software just by describing what they want in natural language. And while that vision is exciting, we’re not there yet. The dream of building complex, production-grade apps without any technical knowledge still remains out of reach.
Technical understanding is still critical, especially when things go off-script — which they almost always do. The biggest hurdle for non-technical users isn’t generating the first version; it’s what comes after: reviewing, modifying, and debugging code. When the AI-generated output doesn’t behave as expected, fixing it still demands a developer’s eye.
That’s why the most meaningful progress right now is in how developers use vibe coding. It’s not replacing them — it’s accelerating them. Using natural language to scaffold applications or prototype features is a massive productivity boost. This is exactly the space Firebase Studio (formerly known as Cursos) is targeting — not replacing devs, but supercharging their first steps. It’s an evolution in how we build, not a revolution that replaces builders.
So, where does all this leave us?
Prompting, image generation, vibe coding — they all point to the same direction: a shift in how we build things. Not a full revolution (yet), but a redefinition of the interface between our ideas and execution. We’re entering a new phase where clear thinking and communication are becoming the new technical skills.
You don’t need to be a developer to start building with AI, but you do need to think like one — breaking things down, being specific, and learning how to iterate. That’s what prompting teaches. That’s what writing a great image brief forces you to do. And that’s why vibe coding isn’t about replacing engineers, it’s about unlocking speed and creativity — especially for those who already know what they’re doing.
In this early AI era, investing in your ability to translate intent into prompts might be the single best decision you can make. Whether you’re building a product, designing an image, or just exploring ideas — how you ask is everything.
We’re not just searching anymore.
We’re shaping.
If you reached this far of the issue, by god, you are amazing!