For 300 years, the Ottoman Empire banned the printing press. The guilds that depended on hand-copied manuscripts had enough influence to hold the line. They protected their profession. They did not protect their clients. When the press finally arrived, it didn't just change how books were made. It changed who got to make things at all. The guilds succeeded in delaying mass literacy in the Ottoman Empire by centuries. Europe colonized the world.
We are a cloud engineer, a dentist, and a biomedical engineer. Without any front-end development experience, we rebuilt the entire Vixul platform. The question for tech services founders is not whether this is happening. It is which side of the press you are on.
Vixul needed a new website. The old one was built on HubSpot — functional until it wasn't. When HubSpot removed features we had been grandfathered into, that was the final push we needed to leave.
The problem was how. We were also mid-repositioning — new messaging, new ICP, new product framing — which meant we couldn't just migrate the old site. We had to rebuild everything while figuring out what everything should say.
We built vixul.com with Claude. Not to assist a developer but as the developer. We described what we wanted. Claude wrote the code. We reviewed, redirected, and iterated. The stack that emerged was AWS: CloudFront, S3, Lambda, and API Gateway, AWS Cognito for authentication, deployed via CloudFormation with a CodePipeline CI/CD pipeline triggered from GitHub and notifications in Slack. Unit and integration testing handled by Playwright. None of this was predetermined. It came from conversation.
I accomplished more playing with my kids over a week than a team of developers in four months.
The first shift wasn't technical. It was psychological.
We had been trained to think about software in terms of what we could and couldn't do. Our skillset defined our ceiling. Vibecoding dissolves that ceiling, but not immediately. The first sessions feel tentative. Your instinct is to ask whether you're doing it right. That instinct is the old mindset reasserting itself.
The shift happens when you stop asking whether you're doing it right and start asking whether the output is right. When we made that shift, the pace changed entirely. We stopped deferring to an imagined expert and started trusting our own judgment about outcomes.
Since 2022 we have been aggregating our operational data in Airtable. We knew we were sitting on a goldmine but couldn't prioritize taking advantage of it — we were too constrained by the limitations of SaaS tools, their layouts, and their integrations.
That constraint is gone. Instead of bending our processes to fit someone else's software, we are now building the right tools for our business. Current proofs of concept include a blog post assistant, carousel creator, video clip generator, brand manager, testimonial finder, and case study generator — all of which we are in the process of operationalizing.
This is what custom unlocks that SaaS never could: tools shaped exactly to how you work, built on top of data you already own.
For two decades, the default answer to any operational problem has been: find the SaaS tool that solves it. The cost of building custom software was too high. So you bought instead.
Vibecoding breaks that logic. When the cost of building custom collapses, the buy vs. build equation changes.
We did not set out to prove this. We just started building. By the time we had finished the first sprint, we had bypassed tools we had been paying for, cancelled subscriptions we had assumed were permanent, and replaced them with software shaped exactly to how we work.
The SaaS tools we bypassed were not bad products. They were simply no longer the most efficient path to the outcome we needed. That is a different kind of pressure than SaaS companies have faced before — not a better competitor, but the elimination of the need for the category.
For tech services founders, this is not an abstract trend. It is happening in your clients' organizations right now.
If a team with no front-end developers can build a production AWS infrastructure in weeks through conversation, the barrier that used to separate technical from non-technical people is gone. The architecture decisions, the infrastructure choices, the system design: these are no longer the exclusive domain of engineers. What vibecoding cannot replace is the business context: knowing what to build, for whom, and why it matters. That is where the value of a tech services firm now lives.
Your clients are sitting on years of unstructured, underused data...and most of them know it. They cannot warehouse it, query it, or build on top of it. That is not a technical problem. It is a strategy problem. And it is the first problem a repositioned tech services firm can now solve, because the tools to act on that data can finally be built to fit the problem, rather than bending the problem to fit the tools.
This is not a reason to panic. It is a reason to reposition. The firms that thrive will not fight to protect what vibecoding is replacing. They will move up into judgment, strategy, and the embedded partnership that tells clients what to build before they know to ask. And the firms that develop that proximity will find something unexpected: when the cost of building collapses, the number of problems worth solving explodes. There are a hundredfold more projects that are now economically viable. The opportunity is not shrinking. It is expanding for the firms positioned to see it.
The guilds that banned the printing press bought themselves 300 years. The human computers who calculated trajectories by hand were replaced in a decade. Your clients won't give you 300 days.
The firms that survive this shift will not be the ones that resisted it longest. They will be the ones that used it earliest to get closer to their clients, to build what their clients couldn't articulate, to become irreplaceable before irreplaceable meant something different.
We are bringing together the founders building those firms at VixulCon on April 20th in Austin, TX. Join us.
Editor's Note: This article was drafted on Tuesday for publishing on Thursday. To give an idea of the pace of developing features, the "What's Next" section was already outdated by the time this article was published.