But recently, I prefer sharing it on my personal page instead, because there are people who slowly start to understand and walk with me on this path: “writing a runtime that runs on cloud infrastructure.”
About 6 months ago, I started sharing about Kitwork and why it was born inside a README file. Many people laughed at it and saw it as something unrealistic. But I don’t understand why, the more that happened, the happier I felt. Because I knew it was right, and there were still a few people who supported me. I’m really grateful for that.
Maybe a few months ago I went back to my wife’s hometown to rest and think about what Kitwork and this journey really are. I kept asking myself why I should build it, and what makes it better than the stack I spent 5 years of my youth building and then replacing.
That question stayed with me for a long time.
Until the first runtimes started to take shape.
And then I began to understand why I am building it.
Just like the first sentence I wrote about Kitwork: “Logic living.”
While many people are still asking: “what’s the difference between microservices and monolith?”
I was asking a different question:
How can logic move across cloud infrastructures?
Maybe in the past, that was impossible, or extremely hard to implement, or even if possible, it came with too many trade-offs.
But then AI arrived.
And human intelligence is no longer bound by simple traditional constraints, which made me start thinking differently about all of this.
I used to ask very naive questions like: why do we need Docker, why do we need node_modules, why do we need k8s, why do we need Redis.
I’m not rejecting technology. I just want to understand the “why” behind it.
Programming doesn’t feel like it used to anymore.
It’s no longer just code and deploy, or infrastructure deploying systems.
Now code becomes a conversation between code, cloud infrastructure, and AI.
AI is slowly penetrating many industries in society. It no longer feels like a “ledger system”, but more like a new notebook that everyone has to start using, just like when we first adopted the internet, then smartphones, and now AI.
I used to know someone who said AI was stupid. Now that same person is using AI every day because they think it has become smart enough.
We are not meant to be attached to any fixed belief. We have to adapt it into our daily life.
Just like how we code, build software, write web apps… vibe coding is fine, AI engineer is fine, loop engineering is fine.
What matters is that we keep learning new things every day to adapt to this world.
I’m not saying you must chase everything. I’m just saying you either slowly adapt, or you get left behind.
And I just hope people don’t say things like: “I just generated a website, developers are going extinct…” please, don’t say that.
And don’t look down on vibe coding that produces a localhost:3000 app, because we still need to learn and grow through it.
If one day I really become “obsolete”, I will probably go back to my hometown, raise chickens, grow vegetables, and live a simpler life.
And don’t be afraid that job titles will disappear, like front-end developer, back-end developer, full-stack developer… new roles will always replace them.
What does not change is often what stops evolving.
It sounds painful, but it’s probably true.
As AI grows rapidly and spreads into every corner of life, localhost on a phone is no longer just a technical story. It becomes a challenge for both people who already know it and those who don’t.
Connecting through a LAN IP and running it on a phone is easy, simple, straightforward.
But actually running a real localhost-like experience on a phone is not just a journey, it is a path full of trials, failures, and retries.
Localhost on a phone makes me think about a new way of programming, where space, devices, ideas, and infrastructure are no longer separated as “developer concerns.”
People often ask me: how can a developer write such long writing like a writer?
So what should I write instead? System design? Bytecode on cloud? Runtime internals? Things you already know?
What I want to write about is not loop engineering or implementation details, because AI can already teach you those, and maybe even better than me.
What I want to talk about is how small things slowly enter and reshape our daily lives.
And why should anyone trust me more than AI anyway?
Multi-tenancy on cloud infrastructure together with AI is what I’m currently exploring and learning every day.
That’s also why I’ve started liking Claude Code and AI agents, because I believe that in a few months, or a few years, we will reach something like autopilot code.
Localhost:3000 on a phone looks simple.
But building it is not as simple as you think.