Between Foundations and Frontiers: Why I write
Exploring fundamentals and frontiers: from protocols to AI agents.
Welcome — this is the first post of my new publication. Here I’ll explore both the foundations of technology (protocols, cryptography, systems) and the frontiers (AI agents, observability, trust). My approach is simple: learn deeply, build something concrete, and write it down. If that resonates, you’re in the right place.
AI agents are the most exciting and fragile systems we’re building today. They can plan, act, and surprise—but they fail in unpredictable ways. I come from a background in systems and quality, where reliability is non-negotiable. When I look at today’s agents, I see what the web looked like before TLS: powerful, chaotic, untrustworthy.
That gap between potential and reality is what draws me here. But agents are not the whole story. My curiosity moves between fundamentals—cryptography, protocols, distributed systems—and frontiers—AI agents, observability, and reliability at scale. These aren’t separate worlds. They reinforce each other.
This publication is my attempt to explore both ends of that spectrum: the roots and the edge. To learn, to build, and to share.
Why I chose this direction
This approach grows out of a few convictions I hold about technology and learning:
Never stop learning. Knowledge decays faster than ever. Curiosity is not optional if you want to stay relevant.
If you can’t build it, you don’t really understand it. Reading or talking about technology is easy. Building is the test of whether you’ve internalized it.
Foundations matter. Protocols, algorithms, and core systems are not boring—they’re leverage. They make mental models sharper and deeper.
Frontier tech opens possibilities. Cutting-edge tools like large language models give us creativity and power that no other moment in history has offered. Ignoring them is wasting the biggest opportunity of our time.
These beliefs push me to move between the bedrock and the frontier. The fundamentals anchor me. The new technology keeps me awake.
Why start with agents
There are many possible frontiers. Why focus on AI agents? Because they sit at a rare intersection: massive opportunity combined with glaring problems.
Opportunities: Agents promise to automate work that was previously impossible—reasoning across domains, interacting with APIs, chaining tasks together. They open new workflows and even new industries.
Problems: Agents fail unpredictably. They hallucinate, loop, or produce outputs that look correct but are dangerously wrong. There is no real standard for observability, contracts, or debugging.
That combination—promise + fragility—reminds me of the internet in its early days. Before TLS, the web was already transformative, but nobody really trusted it for commerce or serious communication. Protocols changed that.
I believe something similar must happen for agents. They need a layer of trust: protocols, observability, and practices that make them usable and dependable. That’s where my systems background meets the AI frontier.
What you’ll find here
This Substack is not a hype feed and not a tutorial farm. It’s a builder’s log. You’ll find:
Essays on fundamentals — cryptography, protocols, distributed systems, and the lessons they hold for today.
Building notes — small demos and prototypes, or even working solutions that solve real problems, with code, traces, and failures included.
Explorations of frontier tech — agents, memory, reasoning frameworks, observability tools.
Reflections and notes — short fragments on what I’m reading, testing, or noticing along the way.
The mix will vary, but the guiding principle is constant: learn → build → write.
What I won’t do
There are already enough blogs promising “10 prompts you need to try” or “Top 5 AI productivity hacks.” That’s not me.
I’m not here to farm clicks or repeat news. I’m here to explore systems seriously, to build real tools, and to articulate how fundamentals and frontier technology fit together.
Where this is going
I don’t see this as a linear roadmap or a fixed theme. It’s an exploration. Sometimes the work will take me deep into the mechanics of protocols and systems, sometimes into experiments with agents, and sometimes into notes from reading and reflection.
The common thread isn’t the specific topic but the approach:
Learn deeply.
Build something concrete.
Write it down and share.
Over time, the connections between these explorations will emerge. That’s the point of documenting them here — to trace how fundamentals and frontiers inform each other.
An invitation
If you care about learning deeply, building reliably, and applying technology creatively, welcome.
You’ll see experiments, failures, successes, and notes along the way. Sometimes I’ll dig into protocols, sometimes into observability, sometimes into the messy reality of building with today’s LLMs.
If you’re also building in this space—whether on the fundamentals side or on the AI frontier—I’d love to hear from you. What’s breaking for you? What have you learned the hard way?
This is not a one-way broadcast. It’s a place to investigate, share, and connect with others who care about the same problems.
Between foundations and frontiers lies the work that actually matters: learning enough of the roots to think clearly, and using that clarity to build at the edge. That’s the journey I’m on, and this Substack is where I’ll document it.