I’m an independent software researcher & writer who explores new ways of doing things, while generally trying to find the best way of doing something, guided by some ineffable sense of elegance.

Because fundamentally,

Ideas are cheap; nowadays methodology matters more than ever. For all the low hanging fruit is gone; the simplest opportunities have been taken. The startup space itself is liken to the first prospectors who got the gold nuggets lying in the riverbed; today the reality is such that you’ve got to dig deep or go home, and those who dig into the depths cannot use mere forethought, for what lies beneath cannot be seen by the mind’s eye.

Just as the early oil prospectors didn’t have precise maps—just intuition and test drills. They had to drill, analyze the pressure, and adjust. The best oil fields weren’t found by foresight alone, but by adaptive exploration.

Just as neural networks themselves are iteratively refined, from a randomly provisioned series of numbers, data scientists apply a methodology in such a manner that the system converges towards some desired outcome, the result of which cannot be perceived upfront from static foresight.

The rise the LLMs like ChatGPT should make it plainly obvious that static foresight has its limits.

In times such as ours, methodology is now more important than ever; success is found only through iterative refinement, not static foresight.

So,

• Choose methodology over static foresight.
• Choose methodology over static business models.
• Choose methodology for uncovering ideas from the very depths where forethought cannot apply.

This is where the researcher archetype comes into play.

As every researcher is essentially uncovering ideas from the depths where forethought cannot apply. Such individuals are naturally curious, organized, and detail oriented. But most of all, such individuals are adept at gathering ideas, taking a hunch, and then digging into the depths—via adaptive exploration—where static foresight cannot apply.

Personally, software research should be applied for developing software that just “feels right”. As a passionate SubStack writer, how I wish things were better!

More generally, there’s so much software in everyday use that just doesn’t “feel right”. Pure and simple, and if these companies don’t innovate by uncovering ideas from the very depths, they will inevitably be superseded by those who do, because the world needs good software!

But while good software improves the lives and productivity of the people. The best software creates whole new ecological roles and niches. (Although personally I just want the software I use to “feel right”.)

(Limited to open source projects.)

I can’t do everything but what I can do I do well.

My philosophy has been to focus on interesting ideas. So far the highlight of my research (in order of coolness) has probably been:

So far the highlight of my research (in order of coolness) has probably been:

1. Publishing hand drawn notes as self contained webpages in a resolution independent manner (why this is particularly cool is because it’s a fully freeform environment as webpages, also the editor includes some brilliant ideas for—among other things—dark mode support that no other note taking app does for some reason).
2. My brute force image compression optimizer (long story)
3. My native (iOS/macOS) markdown renderer that supports all GitHub Flavored Markdown block types with full text-selection support (with multi-cursor support) in terms of TextKit 2 APIs which overall is probably one of the more practical components I’ve worked on (every single iOS app with a chatbot feature needs this).
4. I also once made an indentation based shell scripting language with some much needed improvements that all unix systems need... Although this is probably the least loved of all my projects for some reason. But this is not cool?

   from ./backend/services do
       stack build
       stack exec provision-db
       psql -d services -tc
           """
           CREATE EXTENSION IF NOT EXISTS "uuid-ossp"
           """
   

   Its indentation based shell scripts with higher order functions, multiline strings, and etc. Why is this not a thing?
5. Regarding more abstract problems in computer science, I’ve designed and implemented a very nifty monadic parser combinator library in Swift. The monadic design here is just so elegant because it propagates the parser state between function calls without having to explicitly pass such around. Which just works so simply, and for that I think it’s beautiful.

Also someone needs to make a markdown based spreadsheet app using my native markdown rendering engine because such just seems superior to preexisting options in literally every conceivable manner possible including in terms of a native UI that just “feels right” and iOS users in particular (gotta love them ❤️) have been know to be super picky when it comes to text functionality (just ask the Flutter team.)

Contact me if you need someone to figure out how to freakin implement horizontally scrollable text layout fragments which I’d love to solve once and for all.

Notably, the first project has been most enlightening when it comes to publishing online content.

While I’m still compiling the ultimate list (I've worked on numerous closed source projects that aren’t present), see PROJECTS.md for the most comprehensive list of past projects I've so far worked on, notably with regard to my past open source endeavors.

Moving forward I don’t see myself publishing open source software.

From a business perspective, I don’t believe open source business models are viable.

As a programmer who doesn’t use LLMs I don’t subscribe to any chatbot because everything I do is just too custom. Nowadays we forget that automation only applies to the most common of use cases, which for me is usually within the free tier.

Honestly as a high octane thinker if I had kids I wouldn’t even want them to use LLMs unless I wanted to raise those with mentally inert mindsets that can only think in terms of corporate speak.

If anything I’d love to explore anti-AI themed startup opportunities one day because it’d just be super ironic if such succeeded where all those success chasing startups failed.

• PROJECTS.md: A WIP mega overview of all my major projects. The culmination of my life. Mostly contains my past open source endeavors. Still need to compile a more comprehensive list including my various closed-source projects.
• STATUS-LOG.md: my general timeline of activities.
• QUOTES.md: thought providing quotes on systems.

Local:

• articles/choose-your-constraints-wisely.md: Choose Your Constraints Wisely
• surviving-where-others-fail.md: Surviving where others fail by following the way. Something I wish I knew when I was younger. Perhaps it’ll save you one day.

Substack:

• A Better iOS/MacOS Native Markdown Renderer
• Journal Entry • 2025-2-16: Immigration • Sustainability • Social Cohesion • For non-western countries
• Choose Your Constraints Wisely
• Surviving Where Others Fail
• Honesty Is The Best Propaganda
• The ‘next big thing’ is Social freakin Entrepreneurship!: Chief of all social problems nowadays is the lack of social cohesion. Which is a problem that only bottom-up community building can solve.
• Why does the world need software researchers?: Ideas are cheap; nowadays methodology matters more than ever. For all the low hanging fruit is gone; the simplest opportunities have been taken.

In general these are not yet reading for public consumption.

• The far-future of AI: discusses a future for AI that itself evolves with the progression of time.

  This is a short rough draft that contains many points, all of which I’d like to enumerate on in a series of articles sometime in the future.

  IMO this is a true model for a ‘brain’ that will be the interface between humans and all digital/analog information as it changes with the progression of time. A true model for an assistant will be perpetually active, integrating the diverse array of signals from your data sources, into summary information available at your request, or actively engaging with you when conditions merit your attention.

  Some of my thoughts thereto are inspired by the book Time-Space, Spiking Neural Networks and Brain-Inspired Artificial Intelligence.

  “Everything exists and evolves within time–space and time–space is within everything, from a molecule to the universe. Understanding the complex relationship between time and space has been one of the biggest scientific challenges of all time.”

  Which may be an understatement, this may be the greatest scientific challenge of all time and it’s developed will happen not be via the scientific method but through the ancient engineering method (that built Europe’s cathedrals at a time when the ‘why’ wasn’t understood) and so therefore a new generation of creative individuals are needed, trained to think in terms of problem solving heuristics.

  Just as in computability theory we know there exist problems that are unsolvable, and for such problems we put on our creative hat and think in terms of imperfect heuristics that can at least work in narrow spaces, and through such a manner overcome problems that are otherwise mathematically impossible to solve.

• SubStack: My new preferred place. Publications:

  • Colbyn's Personal Substack
  • NextGen Community Research: WIP! 👉 See the coming soon article

• Medium: Discontinued; the community is dead and now its just AI populated garbage; all thats left is the old stuff from the old me.

• My Beautiful Math Notes: Looks best on wide viewports.

• My Experimental Chemistry Notes: from the root table of contents, click on any option to see my freeform notes.

• Email: hello@colbyn.com