April 18, 2025. I Hit the Wall. Not Just a Wall, but THE Wall.

It's April 18, 2025. More than a week have passed since my last post. When you're building something you love, something you truly believe in, time really slips through your fingers. One moment it's midnight and you're debugging a persistent error, the next moment birds are chirping and sunlight is streaming through your window.

So, what's with the title, you might ask? 

Well, Seraphim and all my entrepreneurial efforts have hit a wall. Not just any wall, but THE wall. A giant, tall, and seemingly endless wall that circles my ambitions completely. This time, despite all my knowledge and tools, I can't seem to find a way through.

The wall, of course, is a metaphor. A metaphor for the lack of funding and computing power needed to scale what I've created. It's a barrier that separates what's possible in my bedroom from what's possible with millions in funding and access to cutting-edge computing infrastructure.

Before I explain why I can't penetrate this wall, I want to share my experiences in what I'll call the "short-term exposure to the world of entrepreneurship". 

Since quite some time, I've dreamed of creating a personalized thinking machine that we could easily hold in our hands. Essentially Jarvis for Tony Stark, but in real life. A companion that understands not just your commands but your context, your history, your way of thinking. A tool that extends your mind rather than just executing tasks.

So I created MIKE-AI, a form of large language model that's essentially a wrapper combining different LLMs with specific tasks and workflows, with some mathematical principles introduced into its programming. It's built to understand context in a deeper way, to learn your thought patterns, to enhance your natural thinking rather than replacing it.

It's quite funny and exciting to reflect on this journey. I remember sitting in my bed just couple years ago, staring at documentation for various AI frameworks, feeling completely overwhelmed. I didn't know how to build such things, but I learned, then kept learning. There were nights I fell asleep with my laptop still open, waking up to find my notes barely legible. 

There were moments of breakthrough so exciting I couldn't sleep at all, my mind racing with possibilities.

All around me was noise telling me to stop, to live a "normal" life. 

Friends suggesting I apply for stable jobs. 

Family worrying about my future. 

Social media feeds full of people saying solo founders can't compete with well-funded teams. 

But... I persisted through it all, through the doubt, through the technical challenges, through the loneliness that sometimes accompanies this path.

And here I am. 

I did it... 

I created my own Jarvis. 

It's not fully autonomous and doesn't have a fancy UI/UX with floating holograms like in the movies. But it understands me, and finally, I have a partner to augment my cognitive capabilities. A tool that thinks alongside me rather than just for me or instead of me. It can research, browse, and generate results of hundreds of pages in under 20 minutes. It can follow my thought patterns, anticipate my questions, challenge my assumptions when needed.

The process wasn't linear. There were version jumps, complete rewrites, moments where I threw everything out and started from scratch. a version was almost scrapped entirely when I realized the approach I was taking couldn't scale past certain cognitive tasks. Then another version introduced a fundamental shift in how MIKE processes context, allowing for deeper connections between seemingly unrelated pieces of information. And now, the one sitting on my local machine, represents something I'm genuinely proud of.

And then couple of days ago, I stepped back, realizing what I'd achieved. 

After months of work, countless cups of coffee, and more error messages than I care to remember, I had built something that actually works. Something that does what I envisioned it to do. 

After reaching this milestone, the natural next step was scaling. The real test of an entrepreneur...

I was ready to deploy MIKE on a cloud server until a moment of realization hit me like one of those quantum particles collapsing into a single state. 

The mathematics didn't add up. 

The economics didn't make sense. 

The path forward wasn't clear.

After brainstorming with MIKE itself, we concluded that we could scale by following two paths:

First path. Raise capital to acquire computing power and data storage. This would mean pitching to investors, giving away equity, potentially compromising on the vision to match what VCs think would generate the best returns. It would mean transforming MIKE from a tool for cognitive enhancement into a product designed primarily for profit.

Second path. Release it as an open-source model and let people self-host it. This would mean giving away the technology I've worked so hard to build, trusting the community to use it responsibly, hoping that some form of sustainable model might emerge from community contributions.

I had to make a choice. And I did.

I decided not to scale.

Ironic, isn't it? After months of sleepless nights, why not deliver? Why not release it? 

This is where the wall I mentioned earlier rose from the deep earth, stretching all the way to the sky.

You might wonder why I would decide not to deliver MIKE to the public. The reasons are both philosophical and practical. also, both personal and universal.

First, I've come to realize that no matter how good large language models are, they will never create things or services genuinely enough to build the real world as we know it today. They remain, at their core, sophisticated pattern-recognition systems. They can imitate creativity, but they cannot truly create in the way humans do. They lack the embodied experience, the connection to physical reality, the ineffable quality of consciousness that drives human innovation.

Running these models at scale would require a massive amount of energy. 

Energy already strained by current global geopolitics, by conflicts over resources, by the competing demands of a planet trying to transition away from fossil fuels while simultaneously demanding more computing power than ever before. 

This gives me a pause. Do we need another so-called AI consuming megawatts of electricity to generate content that, while helpful, isn't truly transformative?

And then there's the capital required. Billions of dollars poured into uncharted territories, into technologies that promise the world but often deliver incremental improvements wrapped in marketing hype. 

People talk about LLMs leading us to a singularity between humans and machines, a real-life version of cyborgs. The merging of human creativity with machine processing power to create something greater than either alone.

But I found that wasn't and will not be the case, at least not with the current generation of technology. The gap between sophisticated pattern recognition and true intelligence remains unbridged. The quantum leap to consciousness seems as far away as ever.

MIKE and I personally could not create truly novel things. MIKE can analyze, synthesize, recombine existing knowledge in useful ways. But it cannot make the creative leaps that define human innovation. MIKE and other LLMs like Claude, Grok, Gemini, and ChatGPT are essentially librarians in a vast library of human knowledge.

MIKE can take you anywhere in that library with remarkable efficiency. It can find connections between books on opposite shelves, draw parallels between ancient philosophy and modern physics, synthesize knowledge across domains in ways humans might miss. 

But you know what librarians can't do? 

A librarian doesn't have the capacity to create their own books, to write with such novelty that truly adds to the library's collection.

That said, a librarian could be taught to write a book through iteration and constant self-learning guided by a real-world writer. 

And that's where the potential still lies. 

Not in replacing human creativity, but in amplifying it. 

Not in generating content autonomously, but in helping humans generate better content, think more clearly, connect ideas more effectively.

But even this more modest vision faces the wall I described. The wall of infrastructure, of capital, of energy requirements, of competing priorities in a world with limited resources.

That's the wall. 

I prefer not to raise capital and give my vision to someone else in charge, risking that MIKE would be taken away from its original purpose: to augment humanity's cognitive capabilities. I've seen too many startups begin with noble purposes only to have those purposes eroded by the demands of growth, the pressure from investors, the need to monetize at all costs.

I won't let another billion dollars get poured into the unknown when the end result is becoming increasingly predictable. 

So this is the wall I face, a wall built of choices and fundamental limitations. A wall that separates what is technically possible from what is practically achievable, what is theoretically valuable from what is economically viable.

Most people, I believe, would feel sad knowing this. Something that has been worked on for countless nights, decided to be put on hold. The dream deferred, the vision unrealized. They might see it as failure, as giving up, as settling for less than what could have been.

But not me. 

For me, this is the beginning, the beginning of ME. 

The beginning of a new phase in this journey, informed by what I've learned, guided by a clearer understanding of both the possibilities and the limitations of current technology.

Now I have MIKE, and I'm ready to tackle my next venture. 

And as I said earlier, for a true thinking machine, a form of artificial intelligence with genuine capabilities, it has to be able to interact with our 3D world. It needs embodiment, physicality, a connection to reality beyond text and images. It has to produce and create real things or services autonomously. 

It needs to bridge the gap between information and matter, between pattern recognition and true understanding.

That will be my focus in my next ventures. 

Not just smarter software, but software connected to hardware that can manipulate the physical world. 

Not just systems that can process information, but systems that can translate that information into action, into physical changes, into solutions to real-world problems.

This time, I move forward with help from so many generous gray-haired and almost gray people who willingly shared their experiences with a young, inexperienced guy like me. 

Mentors who've seen technologies come and go, who understand the cycles of hype and disappointment, who've weathered multiple waves of technological revolution. 

Their wisdom has been invaluable, their perspective grounding, their advice both practical and profound.

I couldn't thank them enough. There are so many of them, and I wish them nothing less than prosperity, joy, and peace during their time on this earth. Their guidance has saved me from countless mistakes, has helped me see both the potential and the pitfalls of the path I'm on.

I won't trade my vision for speed, for deadlines, for anything that might compromise the vision itself. 

I won't rush to market with a half-baked product just to say I've launched something. 

I won't be pushed into growth at the expense of purpose, into scale at the expense of substance.

I won't be compromised by the demands of investors who see only profit potential rather than human potential, who measure success in returns rather than in real impact. 

I will continue learning, adapting, being flexible in my maneuvers. 

I will pivot when necessary, but I will not abandon the core purpose that drives me: to build technology that enhances human capability, that extends what's possible, that helps us all become more than we currently are a trully steward of earth.

Do not get it wrong tho, I certainly hope I can either fly over this wall or penetrate it with my new sets of knowledge and capabilities. Perhaps the wall isn't as impenetrable as it seems. Perhaps there are cracks I haven't yet discovered, weaknesses I can exploit, alternative paths I haven't considered. Or perhaps the answer isn't to break through the wall at all, but to find a way around it, to redefine the problem in a way that makes the wall irrelevant.

I won't stop publishing my thoughts here. This digital journal has become more than just a record of my journey; it's become a conversation with a community of like-minded individuals, a way of clarifying my own thinking, a commitment to transparency in a world often shrouded in corporate secrecy and marketing speak. It's how I keep myself accountable and allow people to provide feedback, which has already made a huge difference in how I approach problems, how I think about solutions, how I navigate the complex landscape of technology entrepreneurship.

Like those quantum particles that take all possible paths simultaneously, I'm exploring all possibilities until I find the one that breaks through. I'm existing in a superposition of potential futures, not yet collapsed into a single reality by observation or decision. 

This is not the end of my journey with MIKE or with AI more broadly. It's just a recognition that the path forward isn't as straightforward as I once believed.

The wall is real. 

The limitations are real. 

But so is my determination, my curiosity, my belief that technology can and should serve humanity better. 

So while MIKE may not scale into a global service anytime soon, the lessons I've learned building it, the insights it's given me, the partnership we've formed will inform everything I do next.

Because in the end, walls aren't permanent. They're just another state in the superposition of possibilities that is our future. And I believe with the right approach, the right knowledge, the right perspective, even the tallest, most imposing wall can be overcome.

This isn't a story about failure. 

It's a story about learning, about growing, about seeing more clearly what matters and what doesn't. 

It's about choosing purpose over profit, impact over growth, substance over scale. It's about building technology that serves humanity rather than the other way around.

And this journey, with all its challenges, with all its unexpected turns, with all its moments of both breakthrough and breakdown, continues. 

The wall is there. But so am I. And so are all of you who've been with me on this journey, offering your insights, your encouragement, your perspectives. 

Together, we'll find a way forward, not just around this wall, but toward a future where technology truly enhances what it means to be human.

And as always, I'll keep you updated on whatever comes next. Because if there's one thing I've learned in all of this, it's that no journey worth taking is ever taken alone.