The last couple of years have been exciting, exhilarating, and utterly transformative. Unless you've been living under a rock, or you are not around technology, you know the feeling—the ground is shifting beneath our feet, and the pace of change, especially in AI, is moving at breakneck speed.
I started programming at 14 and launched my career with my first paid gig at 16, developing software in BASIC. I created many applications for small and medium businesses, until I was about 24, when I transitioned into system integration, networks installations and of course more software development, serving larger and larger clients. From there, I moved into project and program management, eventually leading large enterprise transformation consulting and sales. It's been an incredible journey.
But to understand where we are today, it's worth remembering that AI systems originally started with simple prompts from the command line or scripts. This evolved into the chatbots available to everyone—tools that required users to "learn" how to communicate with them in specific ways to get the best results.
Today, we’re not just talking about the old “prompt engineering” or “can you create this [blank] for me” anymore. While those are important foundations—call that level one—we’re now building so much more. We’re architecting both simple and complex AI-powered systems, crafting intelligent workflows, and embedding AI natively into application stacks. The game has completely changed. And the best part? Every single one of us can do it using plain language.
That's when I realized a profound truth about this new era: you can't just study it. You have to learn by doing, as there are no manuals. This has been my way for a long time. I distinctly remember a time when a University professor spoke in front of all of the class and said: "PCs are a fad that will not last for much longer." There were similar times during the Internet boom, and then again during the mobile revolution.
The old models of learning feel too slow. By the time a course is published, the underlying technology has already evolved. So, I decided to stop preparing and start doing. It has been a blast! This reminds me a lot of the early years of the PC or the internet—that same electric energy of being on the frontier of something revolutionary.
Back to Surfing the Tsunami
My mantra is simple: Start doing, play, make mistake, learn from them, try again! Learn fast, iterate! This wasn't about reckless coding; it was about purposeful, rapid-fire learning through immediate application.
It all started with the foundations: chatbots, LLMs, RAG, fine-tuning, and vector databases. I dove into containers—Docker, Podman, and more—not because a textbook told me to, but because I needed my applications to be portable, consistent across environments, and able to run reliably anywhere. I was juggling multiple programming languages, always choosing the right tool for the job: the steadfast power of PHP, the AI prowess of Python, and the nimble, real-time capabilities of JavaScript, Node.js, and Express.js. Even GitHub, once a simple version control tool, now serves as a sanity check in an environment of continuous updates and relentless iteration.
Orchestrating the Magic
But building the scaffolding was just the beginning. The real magic happened when I started plugging into the AI brain of the world. I began architecting Workflow Automations and designing my own AI Agents. This felt like graduating from being a musician to becoming the conductor of an orchestra. Using different LLMs, adding knowledge by interacting with MCP (Model Context Protocol), Serverless infrastructure utilizing Langbase, and so many other parts of this! SO very exciting! Integration heaven where security and best practices need to be top of mind! (It is so easy to leak a valuable API key if you are not careful!)
This meant getting my hands incredibly dirty with the "behind the scenes" stuff—the APIs. I've been working with secret handshakes to access a multitude of LLMs, connect to text-to-speech services, and integrate real-time transcriptions. It's one thing to use an AI in a chat window; it's another thing entirely to command it through code, to make it a native part of your own application.
The Revelation of "Vibe Coding"
Perhaps the most profound shift in my process has been something I've started calling "Vibe Coding." I've found that AI has become more than just a tool; it's a collaborative partner. It's an incredible, tireless peer-worker that's ready 24/7.
My workflow now is a conversation. I bring the vision, the architecture, the "vibe." I lay out the goal, and my AI partner helps me get there faster than I ever could alone. It handles the boilerplate, suggests elegant solutions to complex problems, and helps me debug in record time. And what's truly astounding is that it's getting better every single week. This synergy is accelerating my ability to build and innovate in ways I couldn't have imagined just a few months ago.
Beyond Vibe Coding: The Age of Independent Agents
But we're already moving beyond collaborative coding into something even more remarkable: independent agents doing complex tasks for you, including entire coding projects and everyday tasks. Today, you can ask an agent to create an entire application from start to finish, or handle your research, manage your calendar, write your emails, analyze your data, and automate your workflows. We're talking about AI that can architect, code, test, and deploy complete solutions while simultaneously managing your business operations, scheduling your meetings, and even handling customer support. These agents don't just help you code—they become your virtual workforce, handling everything from database design to user interface creation to administrative tasks. It's like having a team of expert developers, assistants, and specialists who never sleep, never get tired, and are constantly learning new techniques.
Stop Waiting
This journey has taught me that in the current age of AI, momentum is everything. The sidelines are disappearing. The greatest risk isn't making a mistake; it's standing still.
If you're feeling that itch—that sense that you need to be a part of this revolution—my advice is simple. Don't wait for permission. Don't wait for the perfect course. Pick an idea, make it a project, fire up your editor, and start building something that matters to you, no matter how small it is. You will learn a ton from it!
The future belongs to those who show up! Exciting times!
I have written many articles on my blog (jorgep.com) and on JorgeTechBits Substack but let me know if you need help! Will continue to post my learnings and progress everyone to read.
This has been my journey so far. Feeling very energized and full of curiosity and hunger for learning. Are you feeling the same?
There is No Manual - And As Soon As It's Written, It Changes)