The Industry Decided to Harvest Users. I Decided to Build for Myself.

I started writing code in 1981 on a Commodore VIC-20. I was 11 years old.

Not because I had a plan, not because I saw a career path, I was only 11 after all, but because computers were the most interesting thing I’d ever encountered. I couldn’t stop thinking about them and the endless possibilities that they represented. By the time I was 14 I was writing BBS software and had landed my first paid programming job through the early online community. This was before the internet as most people know it existed. Before the web. Before email was something normal people had. Just modems, phone lines, and a community of people who thought computers were worth staying up until 3am for.

I spent the late 80s and early 90s at the California Air Resources Board and the California EPA Office of the Secretary, supporting their technical operations and going to college part-time at Sierra College. Sierra college had a genuinely good computer science department for the time. I’d sit in the cafeteria between classes with a cup of coffee and sometimes work through problems on a spiral notebook. One of my classmates was a quiet, down-to-earth guy I knew mostly from those cafeteria conversations. One day our professor mentioned almost in passing that this guy had written a tool called FastMerge. Someone he knew was spending hours doing mail merges in WordPerfect. He listened to the problem, spent a week writing a solution in assembler, and handed it to her. The merge took three minutes. He packaged it up and became a millionaire overnight.

Seventeen-year-old me thought that was the most remarkable thing I’d ever heard. Not just because of the money, but because of what it meant. You could solve a real problem for a real person, do it well, and the world would reward you for it. I’ve had a good run being self-employed for the last 32 years. Some very good stretches, some rough patches, but I never stopped believing that was what software, and the tech industry as a whole, was supposed to be.

That’s still the goal. At least it is for me.

Blarg

In May 1994, right after the internet opened up for commercial use, I founded Blarg! Online Services. Purely by accident.

I was working full time at Cellular Technical Services, on contract with McCaw Cellular, doing cellular fraud interdiction work. I’d been playing around with email, Usenet, and UUCP feeds on the side, still half-heartedly trying to get a BBS running. It was mostly a hobby. One day I was talking with a coworker about the internet and he asked me how much it would cost to get access. I told him “$20 a month and a $20 setup fee.” He reached into his wallet and handed me $40. I told him I’d bring him his login information in the morning.

Word spread around the office. Within about a month I had 15 subscribers, 4 modems, and a 56K frame relay connection. It grew rapidly from there.

My wife Jessica and I worked side by side building it for years. She was there for the long nights, the growth, the chaos of scaling a business before anyone had a playbook for it. Blarg grew into hundreds of modems, DSL, shared hosting, multiple DS3s, and eventually became Washington’s second-largest independent Qwest DSL provider.

No off-the-shelf software did what we needed to run the business, so I built TACC, Total Accountability Customer Care, entirely in C++ and Qt. Billing, provisioning, customer support, the whole thing. It even handled incoming support calls, recognizing who was calling and pulling up their account automatically when an agent answered. If you want to see what solo developer software from that era looked like, it’s on GitHub at github.com/gottafixthat/tacc.

Running an ISP at that scale also meant contributing back to the tools that made it possible. I submitted kernel patches for the Cyclades drivers to support the volume of modems we were running, and I’m the person who added MySQL support to FreeRADIUS. Neither of those things made me famous, but they’re out there, quietly doing their job in infrastructure somewhere, which feels about right for how I’ve always worked.

I ran Blarg for about 13 years before selling it in 2007. Jessica had stepped back from the day-to-day a few years earlier. We’d built something real together, and then life moved on to the next thing, the way it always does.

I still have my Blarg email address.

What Came After

After selling Blarg I kept building. Starting around 2005 I began working on a VoIP platform, first on Asterisk and later on FreeSWITCH, focused on business phone systems. I called it CallSphere. It went through several iterations, and after the sale I ended up partnering with the company that acquired Blarg, as they had fiber optic contracts for entire neighborhoods and cities and needed home VoIP to round out their offerings. I focused on that for about ten years.

A few years back they outsourced the VoIP side to a larger competitor, which was fine by me. My interest had been waning anyway. The interesting footnote is that the competitor still hasn’t matched my feature set, so some customers haven’t been migrated yet. I still get royalty checks. They’re not huge, but it’s a satisfying reminder that building things well has a long tail.

Validox came about differently. I was brought in as a consultant to rescue their platform. The company they’d hired to build it had never gotten it working properly. I spent about four months refactoring and rewriting it and got it to where it needed to be. Validox was one of the first companies to seriously tackle compliance reviews in the appraisal space and quickly became the dominant player in the field.

I ended up staying for nearly eight years as their only developer. In client meetings and due diligence processes they introduced me as their CTO, though I was never officially recognized as one, not in title, not in compensation, not in equity. I enjoyed the work and the challenge and was being paid fairly for my time, so I let it go. In hindsight it was a warning sign I should have paid more attention to.

About four years ago the partnership dynamic at Validox became toxic. One of the partners and I had been pushing to expand into adjacent territory in the mortgage compliance space. The other partners weren’t interested. So we left and built it ourselves.

That became Vueterix.

We started with a single product and actively listened to clients about their pain points. Every product we’ve added since has come directly from those conversations. Vueterix has doubled year over year for four consecutive years and is on the cutting edge of appraisal analysis, QC, and AI insights. In an amusing twist of irony, we’ve never sold a single one of our original products.

The platform started as a PHP monolith, but as we grew we started hitting its limits. I rewrote the entire thing from scratch in Go and Quasar. I lead all the technology development.

It’s the same way I’ve always worked.

The Book That Mattered More Than the Platform

I also spent about six years building dittos.io, a communication platform for schools and PTOs. Think ParentSquare but actually usable. It never took off the way I hoped it would, but it did something I’m proud of.

My son was telling us he couldn’t read. But he could. He could read all of his sight words without a problem, but he “couldn’t read a real book.” Challenge accepted. My wife wrote a children’s book built solely around the 70 sight words his class learned in kindergarten, and we published it through dittos.io and gave it away.

We’ve given away close to 8,000 copies to local kindergartens over the last seven years. The platform may not have worked out, but the book did. We just delivered almost 100 books last week, in fact.

Life Gets Complicated

In 2002, Jessica was diagnosed with Multiple Sclerosis. She’s needed mobility assistance for the last ten years and has been in a wheelchair for the last eight. About five years ago we found out she also has a heart condition. It hasn’t been easy, but it’s just the way our life is. We don’t think of it as abnormal because it isn’t, it’s just our normal.

What it does is make certain things very concrete that might otherwise stay abstract. Health tracking isn’t a market segment to me. It’s a daily reality. Last summer, my doctor told me to start tracking my blood pressure and weight and bring the numbers to my next appointment. I went looking for an app that would let me do that cleanly and generate a report I could hand to him. I couldn’t find one I liked, so I built Caffeinated Vitals.

That’s the pattern of my entire career. The tools I needed didn’t exist, so I built them.

What Changed?

I started in a time when computers were tools. They were supposed to make things easier, faster, better. You solved a real problem, you got paid, everyone went home happy. FastMerge. Three minutes instead of hours. That was the deal.

Somewhere along the way the deal changed, and it didn’t happen all at once. It happened the way the frog in the pot doesn’t notice the water heating up. The model quietly shifted from “solve a problem, get paid” to “give it away free, harvest the user.” Search keywords and online purchases became advertising profiles. Tracking pixels spread across the web. Cross-site cookies fed behavioral data back to platforms. Everything that was free turned out to have a cost, just not one paid in money.

The moment that crystallized it for me was when Google quietly buried “don’t be evil” in the fine print of their code of conduct. That wasn’t a formatting change. That was them deciding they didn’t need to lead with it anymore. Most people didn’t notice, or didn’t care. They wanted the dopamine hit of the free internet and were willing to pay for it with everything else.

Today, the focus seems to be on pushing out something as fast as you can, make it look pretty, and sell it before the new SaaS smell wears off. Nobody seems to build software to last anymore.

I find it genuinely disgusting. Not in a self-righteous way, but in the way you feel when something you loved got ruined. The early internet was a community. It was people sharing knowledge and solving problems for each other. What it became is something else entirely.

Why Caffeinated Softworks Exists

I spent most of the last 15 years B2B. The work was good, the clients were real, but nobody knew my name. That was expected in that world. You built things, other people took the credit, and you moved on to the next thing. I was comfortable with that for a long time.

I’m less comfortable with it now.

Caffeinated Softworks started as a holding company to consolidate my various ventures. But when I started building consumer apps, something shifted. I was building tools I actually used every day. My wife was using them. I was solving problems I actually had. The Checkbook app exists because I fell off the GnuCash wagon during a chaotic few years of moves and medical appointments and never found anything that synced cleanly between my phone and my desktop. Notes exists because Joplin didn’t pass the WAT (the Wife Acceptance Test), and Notestation kept losing its authentication. Calendar exists because I was tired of seeing my wife’s medication reminders cluttering my work calendar and no app would let me filter them without deleting them or removing her calendar entirely.

These aren’t products I designed for a market segment. They’re tools I built because I needed them, and then realized other people probably needed them too.

The philosophy is simple: software should do what it says, respect your data, and stay out of your way. No ads, no tracking, no AI, no accounts required to use the basic features. Your data lives on your device. The subscription is just for sync, because keeping your notes and finances in one place across your phone and your desktop is genuinely useful and worth paying for.

I’m now in my mid-50s and I still write code every day. I still think building things is the most interesting work there is, it’s genuinely my happy place. I still believe there’s room for software that’s honest and useful and doesn’t treat the person using it as a resource to be harvested.

Caffeinated Softworks is my attempt to prove that.

If you want to see what I’ve built, start at caffeinatedsoftworks.com.