I remember the first time I touched the terminal. I was 13. I’d had a PC at home since I was 9, but I mostly used it to play games and talk to strangers on the internet (what could possibly go wrong?)
I was in my second year of ESO (Spanish secondary school) and Monster Hunter had become all the rage. A lot of us had a PSP, and we’d spend our time (…) I found out that a new Monster Hunter had come out in Japan and that it might never make it to Europe, but digging around I discovered there was a patch to translate it into Spanish.
Mind you, I had to open the terminal. I felt like a genuine hacker. A couple of commands and I had the ISO.
From the moment I discovered the Pandora battery, I made a business of jailbreaking PSPs — first for my friends, and later for kind strangers who offered me money. By then I already knew how to run a UMD dump (…)
Today, many years later, I still feel that same fascination with tinkering, but from a very different angle. I’m a designer, not an engineer (though opinions vary), and my relationship with code is more of a platonic love: I adore what it does, but I lack the discipline to learn JavaScript or Python… for now.
And yet, my home has turned into a smart laboratory, and the most surprising part is that I haven’t written a single line of code myself. It’s all been possible thanks to a mix of hardware, software and, crucially, the tireless help of an AI.
The Heart of My Smart Home: Raspberry Pi and Home Assistant
My new Raspberry Pi 4 B is the brain of my home. With Home Assistant (HA), I’ve managed to centralize control of almost everything. My Tapo bulbs and plugs, along with the Broadlink RM4 Mini (which handles my air conditioning via infrared), are now connected and accessible through HomeKit. This means I can control the climate with Siri, my Apple Watch, or from the HomePod when I don’t have my devices at hand. The main idea is to avoid waking my newborn daughter (Sabina) and make sure she never gets too hot at home.
The experience with Broadlink’s original app is a good example of why HA is so crucial: its UX is frustrating. It’s far simpler and more efficient to set up and manage the RM4 Mini directly from Home Assistant, where I have a unified, fluid interface.
Soon, the RM4 Pro will join the party to control my fan, since it uses radio frequency instead of infrared. I’ve also integrated the Tapo H100 hub with its motion sensor, adding another layer of automation and security.
The Huawei inverter for my solar panels is also connected to HA. This integration is essential, not only for monitoring energy production and consumption in real time, but for future automations I’m already looking into:
- Smart blind and curtain openers: To optimize sun exposure or privacy.
- Bluetooth “Finger”: A tiny robot that presses physical buttons. I’d like to use it to start the dishwasher or the washing machine automatically based on the solar production and consumption data the inverter reports. Imagine the savings and the efficiency.
- The idea that electric cars could serve as a bidirectional battery for the home strikes me as the next level of energy self-sufficiency.
The Touchscreen Experiment: ESP32-S3-Touch-AMOLED-2.16
Related to [[ESP32]]
Another fascinating device in my lab is the ESP32-S3-Touch-AMOLED-2.16, made by Waveshare. I bought it intrigued by its possibilities and without knowing exactly who made it (thanks, Hermes, for the research!). As part of an experiment, I’m feeding it energy data from Home Assistant, using it as a small informational display. Who knows, maybe one day I’ll end up buying a tablet for the kitchen for a bigger display and more intuitive control. This device is so interesting that I’ll actually dedicate a full post to it in the future to dive into its uses and tricks.
Beyond the Smart Home: The Raspberry Pi as an Entertainment and Retro Gaming Hub
The Raspberry Pi’s versatility doesn’t stop at home automation. It’s also my media center with Kodi, progressively replacing subscriptions and, in some cases, Plex. It lets me access my content locally and in an organized way.
What’s more, I’ve returned to my gamer roots with a retro gaming and Retro Achievements setup. It’s a delight to relive classics and unlock achievements on consoles of old. This has led me to collect Anbernic handhelds, reaffirming a philosophy I once struggled to understand: “one function, one device.” Years ago, my mindset was “one device for all” (especially before Android had a decent desktop mode), but now I value the dedicated, optimized experience that specific devices like the Anbernics offer.
AI as a Partner: Two Hermes Agents and “Deskilling”
Perhaps one of the most “meta” implementations in my home lab is the AI integration. Right now I have two Hermes agents: one on my Mac for general tasks and another on the Raspberry Pi, dedicated exclusively to managing the house and communicating with my wife, Marta, through our Telegram group. It’s fascinating how a dedicated device (the Raspi) can host an agent with such a specific context.
A practical example of this is how we use the ESP32 to log our dog’s medication. Hermes monitors this data and lets us know when it’s running low, reminding us to place another order. Goodbye to forgetfulness!
Another successful project has been the babycam. We use an old Android phone, set it up in 5 minutes, and HA sends us critical notifications if it detects movement in Sabina’s room. It’s a private, free, and efficient solution for our peace of mind. We still need to integrate the NVIDIA Spark, but that’ll come.
But the most revealing part of this whole journey is the reflection on the role of AI. I’d like the cursor history of this conversation to serve as proof: everything I’ve built hasn’t required me to write a single line of code. AI has been my co-pilot, my translator of ideas into executable commands.
This brings me to a crucial reflection: AI “deskilling”, or how artificial intelligence can lead us to lose certain technical skills. It’s a double-edged sword. On one hand, it democratizes tinkering and creation, allowing people like me, without a formal technical background, to build complex systems. On the other, it underscores the importance of understanding what you’re doing. In my case, the next step isn’t just asking the AI to create a new automation, but reverse-engineering what I’ve already built. Taking the commands apart, understanding the logic, seeing how the dots connect. It’s my way of learning Python and JavaScript without the traditional “discipline”, exploring and understanding what’s already up and running.
This home lab isn’t just a set of connected devices; it’s an extension of my curiosity, a manifestation of how technology, with the right help, can empower anyone to build the future of their home, even without knowing how to code. And yes, one day I’ll find that discipline.
[!NOTE] Editorial brief Idea: “designer without code” — smart home built with AI (Raspberry Pi + Home Assistant + solar + ESP32 screen) + reflection on AI deskilling. Series: part 1 of 2 → links to
ESP32(the Waveshare AMOLED device). Audience: tech + HM + designers; proves you ship with AI. 🔵 AI·work·skill (proof post). Status: rich + readable; trim the “deskilling” section so it lands as a thesis, not an aside (ties toletter-to-20).