Why I'm not a Full-stack Dev
The answer to the question in the title isn't "because I'm a poet" or "because I'm a zookeeper". I am a web developer, but only front-end. And I want to explain why.
I've been programming for fifteen years now. I got started making little Ruby projects in Codecademy and uploading them to GitHub. I quickly learned about Jekyll, one of the most popular static site generators, dubbed "blogging for hackers" by its creator Tom Preston-Werner (GitHub co-founder) in 2008, and fell in love.
Sadly, I decided to delete near-everything pre-2019, so it's difficult to find any of this. A good example of my (bad) code would be one of my old tumblr blogs, Pinedraft, which I wrote the theme for, from scratch. Or my archive of old projects.
I restarted my public-facing dev journey in late 2019. I was let go at my job at the children's hospice due to company changes, and I participated in various employment programs. I was lucky enough to hear about InceptionU purely by chance.
Back then, they had government-assistance for funding, so I could participate despite the ~$19,000 tuition fees. My cohort, the 4th ever (they're now on cohort 15) began in February 2020.
Yeah.
So, the majority of my six months of learning were not in-person at the gorgeous, fish-shaped Central Calgary Public library. Instead, I was working from home, reckoning with quarantine and a global pandemic while trying to learn full-stack web development.
Now, obviously there are spaces in web development where a full-stack solution is needed. And I was fortunate to learn the MERN stack and Django among other tools, stacks, and frameworks during my time learning with InceptionU, but we need to collectively take a step back.
Decades ago, the Internet used to only have full-stack. You either understood the complexities of database management or you didn't have something functioning. Even simple blogs were written with MySQL and PHP (in the LAMP stack).
I don't need to tell you that things have changed. Not substantially, but entirely. There are people that invested time (and money, a lot of money) in Web3 and blockchain as tools and methodologies as a new paradigm for how we should tackle the Internet and its development. I mean hell, the name literally inferred a new version of the Internet.
That never came to pass.
90-95% of Web3 projects have failed, with over $100 billion invested into the space yielding minimal returns. NFT art trading collapsed 93%, from $2.9 billion in 2021 to $23 million in Q1 2025. The Sandbox and Decentraland, billion-dollar valuations with fewer than 100 daily active users each. Seventeen Web3 games shut down in 2025 alone. Google search interest in Web3 is essentially gone since its late 2021/early 2022 peak.
Now, in its place, we have AI-driven or even AI-first web development. The tech stack arguably irrelevant and perhaps unknown to the unassuming vibe coder that simply talks to an LLM token-generating agent that does all the work of actual development.
"Vibe coding," coined by AI researcher Andrej Karpathy in February 2025, is slang for where "developers" use plain English to describe what they want and AI generates the code. By March of this year, Y Combinator found 25% of startups in Winter 2025 had codebases that were 95% AI-generated. A Stanford University study found employment for software developers aged 22-25 fell 20% from 2022 to 2025, coinciding with the rise of AI-powered coding tools.
GPT-5 generates "a larger and more complex volume of code than any other model," making it "a serious challenge to review and maintain." The code is bloated, difficult to understand, and prone to security vulnerabilities. 170 out of 1,645 Lovable-created web applications had security issues that would allow personal information to be accessed by anyone.
Hammers looking for nails. Complexity created because it is grandfathered in, and people making money by prolonging and overcomplicating solutions to problems. This has always been the gold standard of enterprise software.
The truth of the matter, though, is the Internet and creating on it can be far more simple, and counterintuitively, far more permanent and robust. Plain text. Markdown files. Add some APIs and JavaScript and you have the JAMstack, which I've been evangelical about since forever.
Remember Jekyll, that I talked about earlier? It creates static sites. You only need to know how to edit plain-text files in order to make a blog post. Zero databases.
This certainly doesn't mean you don't have technical knowledge or development skills. The website still needs to be designed (front-end), along with hosting solutions, documentation, etc. Overall, though, it is far simpler in a way that works. Performance and accessibility are far higher than expensive, bloated solutions. The evidence is overwhelming:
- According to Netlify's 2024 survey, 44% of devs believe JAMstack enhances site performance by 2x
- Over 60% of new web projects now use a JAMstack or static-first approach, up from only 27% in 2020
- Nike reduced server costs by 40% and improved page load times by 60% after adopting JAMstack
- Over half of users abandon a website if it takes more than 3 seconds to load—JAMstack's pre-rendered content ensures ultra-fast page loads
- Faster websites experience a 70% increase in visitor retention
Most people are creating information online. The exchange of ideas is why the Internet exists in the first place, and ideas can just be plain text. They should be plain text because it outlasts all other file formats or proprietary extension types.
Digital preservation experts consistently recommend plain text and open formats like UTF-8, .txt, and Markdown for long-term archiving. The National Archives holds over one billion files representing more than 700 file format versions, the vast majority being plain ASCII text, HTML, and open formats because of their longevity. As software updates and companies close, proprietary formats become obsolete and unreadable. Plain text will always remain readable so long as there is a computer to read it.
Creating the JAMstack Internet
All of the above is why I started 🍓 Berry House. I want to help people start their websites and online presence, but after that? I want to be able to let go, to let them take the driver's wheel and not need people like me anymore.
The goal is digital autonomy. When you build with JAMstack tools like Eleventy, you can move your site anywhere. No vendor lock-in. No proprietary databases. No monthly subscriptions just to keep your content accessible. Your writing lives in Markdown files that will still be readable in 50 years, whether or not any particular company still exists.
This is the IndieWeb philosophy in practice. Own your platform, control your content, and build for longevity. It's the opposite of the current trajectory toward AI-generated sites hosted on platforms that could disappear tomorrow, taking your content with them.
Berry House operates on a dual mission. We provide professional services to corporations that can pay, while offering pro bono and pay-what-you-can support to marginalized communities and nonprofits. Every paid project funds the ability to help someone who needs a website but can't afford the typical (bloated) $5,000+ agency rates.
More importantly, we build calm, sustainable websites that don't require a developer on retainer. We teach you how it works, document everything, and then step back.
The Internet should be a place where ideas can live and spread without intermediaries extracting rent at every turn. Where an emerging nonprofit in Calgary can publish their mission statement without paying Squarespace $300/year. Where a poet can share their work without feeding it into Meta's engagement. Plain text files and simple pipelines will always outlast the hype cycles and bubble economics.
That's what being a front-end developer means to me. It's vanilla HTML and CSS and JavaScript, sure, but it's also building toward a web that's accessible, portable, and genuinely owned by the people creating it. Not a full-stack developer chasing the next framework or database architecture, but someone focused on the foundational layer: helping people communicate clearly and build platforms they control.