Small progress today.
Moved my ESP32-S3 AI desktop pet onto a dual-breadboard setup, learned to check connections and voltage with a multimeter, and got its OLED eyes and Wi-Fi chat working again.
Next: microphone + speaker.
Building toward my first MVP on Dec22
#ESP32#AIoT
A small milestone in my ESP32 journey.
Today, my little OLED screen finally came to life—with two expressive eyes. 👀
It's just a few pixels, but somehow, it already feels like a tiny companion.
One step closer to building my first AI pet. 🤖❤️
#ESP32#BuildInPublic#AIoT
Claude Fable 5.2 is another reminder: writing code is getting easier.
But building useful software still requires understanding real problems, making good decisions, and delivering reliable results.
The real advantage isn't just coding faster. It's knowing what to build.
A new project we're exploring: managing 50 Android phones from one PC using ADB, USB hubs, and a local dashboard.
Not every useful automation needs fancy AI.
The real skill is identifying problems and building practical solutions.
Still in the planning stage. #BuildInPublic
@elonmusk Three words that, in my view, help explain the Republicans' midterm defeat:
**WAR. IRAN. DOGE.**
So,At least we won't have to see a third mushroom cloud rise.
Three words that, in my view, help explain the Republicans' midterm defeat:
War. Iran. DOGE.
So,At least we won't have to see a third mushroom cloud rise.
AI is making software easier to build.
So what are AI resellers actually selling?
If the business makes more money recruiting agents than serving real customers, that's a red flag.
Validate customer demand before paying for an AI franchise.
Starting my ESP32 journey today.
I have experience building software products with AI agents, but almost zero experience with electronics.
My first challenge? Learning how to solder.
The long-term goal: connect AI agents to the physical world.
I'll share what I learn.
I've been experimenting with AI agents to automate real business workflows.
From researching market needs to building and testing solutions, I'm exploring what a small, AI-powered team can actually achieve.
I'll share what works, what fails, and what I learn.
AI lowered the cost to enter the game, but drastically raised the bar to win.
It gives you infinite ammo, but victory still depends on judgment and human insight.
If you're building with AI today, ask yourself: Strip away the generated code—what irreplaceable value is left?
AI makes building products trivial, but creating true value harder than ever. 🧵
Everyone is celebrating "Vibe Coding". You can ship an MVP in a weekend. Tech is democratized.
But here’s the reality: AI lowered the bar to BUILD, but not the bar to JUDGE.
Here’s why👇
Supply is exploding with fake demand.Easy dev means a flood of AI wrappers built because we can, not because markets need them. Cheap code solving fake problems is still digital garbage AI solves "How", not "What" or "Who".It can’t sit in the room to feel subtle industry friction
When code becomes a commodity, the moats shift ❌ Technical execution & feature bloat ✅ Deep domain intuition ✅ Seamless workflow integration (delivering outcomes, not tools) ✅ Distribution & Trust
In a world where anyone can build a chair, value shifts to taste and ergonomics