Independent software researcher and problem solver with a calibrated sense for what works.
I build software guided by taste, pattern recognition, and an internal sense of coherence. Not everything that works can be made legible, some systems just feel right. That’s the edge.
Available for collaboration or bespoke software systems that prioritize function, feel, and internal coherence.
See the new WIP Timeline Showcase of all my past work.
I'm an independent software researcher and writer focused on identifying and executing new methodologies for building software that just "feels right."
In the modern landscape, ideas are cheap. Methodology is leverage. The age of obvious wins is over and what remains are complex, layered problems that resist surface-level solutions. As in deep sea oil prospecting, success now comes not from foresight alone, but from adaptive exploration.
Much like how neural networks evolve from randomness into purpose through iteration, software innovation today demands the same mindset. The emergence of tools like ChatGPT makes it plain: static planning is obsolete. You either iterate or become irrelevant.
Researchers operate in the dark. They chase what can't yet be mapped. They don't guess they probe, iterate, and refine. The best of them are curious, disciplined, and exacting. That is the archetype I embody.
I write because I want the software I use to feel right. Most of it doesn't. Most of it is designed around business models, not experience. But truly great software does more than just "work" it changes the way people think and operate.
Success is easy. That's its problem. Depth has a cost, those unwilling to pay it should stay on the surface. I’m guided by a different metric: elegance.
As for tech bros: they optimize for the wrong variables. Web clients everywhere. Native software stagnates. Everyone's chasing the obvious, no one's chasing feel. We need new primitives, not more frameworks.
Freeform, resolution-independent publishing with novel dark mode support.
Unorthodox but effective. Long story.
GitHub Flavored Markdown, multi-cursor, native text selection via TextKit 2.
Cleaner syntax, multiline support, higher-order functions.
Parser state propagated implicitly; elegant and clean.
Someone should build a Markdown-based spreadsheet app using my renderer. Native UI, scrollable layout fragments — it would outperform every existing option.
This is not a portfolio. It’s a declaration: I don’t follow systems. I build better ones.
The age of foresight is over.
When the gold lay in riverbeds, vision alone could win. Not anymore. Now, you dig or you get nothing.
Modern systems are too complex, too layered, too emergent. The best opportunities are buried, not obvious. And they cannot be reached through ideation alone. They require feedback, adaptation, iteration. They require exploration.
Consider early oil prospectors. No maps. Just drills. They took action, measured pressure, and adjusted course. The biggest discoveries weren’t predicted they were uncovered.
Neural networks follow the same path. Random values, refined through feedback, converge toward solutions no mind could design from scratch. The process that led to LLMs like ChatGPT proves this: the end state wasn’t known upfront. It was discovered through method, not vision.
Static foresight cannot win this game. Methodology can.
Thinking, if it’s not embedded in feedback loops, is speculation. Execution without iteration is waste.
If you're discontent then do something. But make sure it’s something the mind can’t fully control. Exploration requires a willingness to move without guarantees, to act before you're ready, and to embrace what your competition can’t: the unknown.
Real builders don't follow maps. They build compasses.