Initialized

There is nothing in this world but axioms and systems composed of them. In every detail, and in every engineering decision, exists a particular and specific need, and every need, while different, is composed from similar constraints: cost, time, effort, knowledge, will. Every decision, every axiom, every system is a result of those constraints, and from those axiomatic constraints we get Engineering!

My name is Brett Parson, and this site is my personal canvas for my work, my writing, and a home for a bunch of stuff you lot will likely never see. I am a software developer by trade and, with the advent of Agentic Development, a software developer with a progressively expanding number of Claude Code instances open. My primary interests are web technologies, and I consider myself a web technologist, and man is today an interesting day to be a web technologist!

The World of Software has Gotten Smaller

I take an enormous amount of pride in my work, and have strong opinions on software development. I believe in a free and open web perpetuated not by opaque platforms but rather by individuals and software artisans working together across one of the most efficient and effective communications platforms to have ever been conceived: HTTP. Hypermedia is one of the most important technologies ever designed, and not only is it easy to use, it is also incredibly efficient for global transmission!

Structured text and transportation protocols predate the advent of Large Language Models by a mere few decades, giving us adequate time to construct a training corpus that spans the breadth of human knowledge! Not only are Large Language Models fantastically adept at communicating in the web's native Hypertext Markup Language, they are also perfectly suited to designing systems for serving these templates statically. This happy coincidence means the barrier to entry for hosting and designing web applications cheaply with minimal dependencies has dropped to zero.

Software should be small, fast, and easy to use. Prior to the advent of silicon intelligence, we as engineers were constrained by our time and costs. Because software was expensive, we couldn't afford to repeatedly reconstruct the same systems for new projects, so we relied on the collective needs of the common engineer to determine the common solutions to our common problems. Open Source Software, Frameworks, code designed not for me but for we. This ultimately bloated our dependency trees, and required dependency trees to support our dependency trees (whereby my arch-nemesis npm was born).

hear me clearly, I don't intend to denigrate the efforts of my peers and colleagues here, I emphasize the point that we no longer need be subject to the tyranny of bloat. Custom software and maintenance are cheap! Not free, of course, but measured in pennies and seconds, as opposed to dollars and minutes. The Cost—our first, engineering constraint—has gotten so much smaller than it used to be, and our abilities to communicate have become far more personable, individual, discrete, and powerful.

"But Brett, doesn't your beatific vision of personal computing involve renting gigawatts of compute to write HTML?"

Fair reader, you present such interesting arguments for me to deconstruct! Thank you! While yes, I present what appears at first a contradiction, so too does your current open source software rely on the efforts of the globogigacorp! Software, up until now, has been gatekept by the people who are paid to need it! The vast majority of Open Source Software exists because software engineers who make hundreds of thousands of dollars a year need it and are provided time during their day at globogigacorp to help maintain external dependencies so their internal systems can continue to churn. The layman is hopelessly left in the dark, unable to efficiently organize his knowledge and understanding in his spare time because he can neither afford the time to read dense, technical documentation nor does he care to try. His world is already downstream of the priorities of globogigacorp, and is awash in the sea of competing needs, none of which spend a cent on the concerns of the layman!

By engaging with the globogigacorp's Large Language Model, which they offer for the fraction of the cost of an actual engineer, the layman can efficiently navigate not just the Open Source Software Ecosystem but the entirety of All Software Ever Written (well, mostly)! The inversion of cost, and the democratization of knowledge, has given us a once-in-a-lifetime opportunity to participate in ways never before available to the average webizen!

Systems For Everyone

I've always felt like build-versus-buy was over emphasized. Buy makes sense from a business perspective in some cases where "reinventing the wheel" would cost more than paying someone a monthly fee to solve it for you. Engineering hours are expensive! If you can utilize somebody's engineers, support-time, and complexity to accomplish with 200 bucks a month what would potentially cost you tens of thousands in engineer hours for back-side processes, you just straightforwardly do it. Even with AI, the infrastructure costs of maintenance for cost-centers as opposed to profit centers are plain.

That is, until you introduce Chinese Token Pricing and Open Source Software LLM Harnesses. Most of the effort ends up in defining and describing the requirements in concrete terms to the effect where composing systems is no longer limited to folks who understand how to write a for-loop. Optimal systems are proprietary, private, owned. Solving everyone's problem is a solution for expensive engineering hours, but solving your problems is the optimal solution and optimal usage of engineering hours, provided you know what your problems are. The acquisition and usage of technical knowledge is no longer kept behind a four year degree and 10,000 hours spent figuring things out the hard way; knowledge as a component of the engineer's constraint falls away.

While this is all very interesting from a business perspective, it's even more interesting personally! The "buy-side" of the buy-vs-build for personal/private use is, typically, resolved with "get nothing at all!" Now, however, we don't need to concern ourselves with expensive SaaS implementations solving for enterprise users with freemium models because we can spend ~$2 once with a cheap open source model to build it entirely in an afternoon and own our software directly! This is the most exciting time to be a hobbyist because we get to tinker directly with the software, design our own systems from scratch, and play with the fundamentals for our own purposes to build personal, bespoke software! The effort of building and maintaining has collapsed into a few sentences passed through to a terminal.

While we could have always done this, the limitation has mostly been time-cost and activation-energy. I could build my own personal implementation of Redis by hand, but I have more important things to do regarding the actual content where it becomes unreasonable to consider unless I really wanted to write my own. Now? I'm able to spin up multiple implementations with multiple different languages just for giggles to see how it works, why certain things are built the ways they are, and to tinker with the concepts directly from a high level. While I would never do this at work, I will absolutely do it for myself in order to learn how to orchestrate these systems and build things for myself! Systems For Everyone is my practical exposition of building software with agents for fun, which is the ultimate purpose of this website and blog.

The final gate is, in truth, not a gate! The final constraint of the engineer is the will to pursue the ends. Will is the final, practical constraint left preventing the average individual from building software. Just the intention of seeking the result, of engaging with it, yearning for it, that is the last component that cannot be solved for with large language models. Nothing anyone can build can drive you, the individual, to accomplishing anything without you willing it to be accomplished. Systems For Everyone doesn't mean everyone can host and own their own stack, rather it implies anyone can come to it on their own terms and build it for themselves in a way that was never before realistic or possible.

Final Thoughts For Readers

This blog will serve multiple purposes, and is as of writing an unfinished product for myself. I have a lot of work to do yet regarding the system I'm writing this all in, but in the spirit of agile agent orchestration this is my Minimum Viable Product for my primary use case: teaching and sharing my experiences! Over time I intend for this website to flower into a monument of personal computing and web design, and an example for those interested in digital sovereignty and data autonomy! As a consequence, this blog will eventually be partitioned into different topics, and more concrete books or tutorial pages will spawn once I can support them, as well as twitter-like "notes" that are shorter thoughtforms than a whole blog.

Of all things, I believe thoughtful intention in the details, even in the details nobody will ever see, is the most important thing for any craftsman, regardless of the tools or skills involved. A pretty tool with no use is merely decoration. I'm not interested in mere decoration. If you aren't either, you'll probably find the things I have to say interesting, and I invite you to continue reading and engaging with the site.

Thanks for reading! Eventually I intend on producing an Email subscription through the site, as well as other feeds like RSS, but I have work to complete yet so bookmark and check in periodically. You might find another entry!