003: Jankees van Woezik - Omarchy Plugins and Indie Maker

Jankees van Woezik, a developer from the Netherlands, recounts his early career beginning around 1999, initially creating websites for local punk bands and exploring Linux while studying system administration. He quickly realized system admin work was not for him, especially managing printers, and shifted towards scripting and programming, developing an interest in automation and Linux. Later, he switched to multimedia studies, earning a degree, and has been self-employed since 2007, nearly two decades. Jankees shares his early adoption of Rails in 2006, following Basecamp and 37signals, and his first contact with Omarchy, which he tried initially but didn't connect with until later. He started actively using Omarchy after March 2025, influenced by David's efforts, and transitioned his web development to Rails last year. His journey reflects a persistent pursuit of coding and automation, and a turn towards creating plugins for Omarchy that focus on productivity and automation, which he eagerly shares on social media.

Jankees discusses his early attempts to develop plugins, emphasizing how Quattro's release with a new plugin system and the quick shell component revolutionized his approach. His initial project was a simple toggle for lights in his office, which he developed after a few prompts using AI, and from there, he became addicted, creating multiple plugins and posting them daily. He admits he hasn't looked at the code of his plugins, relying on thorough testing and AI models analyzing for security, as he prefers to focus on the functionality. He highlights how easy it is to develop plugins now, even for non-sophisticated tasks, and notes that the most crucial aspect for him is iterating and building tools that solve his personal needs. His saturation with the Omarchy plugin ecosystem started early, driven by a desire to scratch his own itch, demonstrating a strong DIY ethos combined with AI-assisted development.

Jankees explains that he rarely examines the code of his plugins because they are small and interact with known CLI tools, which he trusts. Instead, he tests plugins extensively in virtual machines and uses multiple AI models to review code for security issues. He mentions that some community members criticize his plugins for security flaws, but he argues they are simple pieces interacting with existing systems and therefore pose limited risk. He discusses his approach to security research, utilizing sophisticated reports and proof-of-concept exploits, running these tests on dedicated VMs for safety. He describes a workflow where vulnerabilities are identified and fixed in isolated environments, using AI to generate patches. He shows a desire to continue innovating on this front, keeping security a priority while automating responses to vulnerabilities.

Jankees reflects on the rapid advancement of AI and its potential to change software engineering fundamentally. He predicts that in the future, coding by humans might become obsolete as AI surpasses human capabilities in writing code, which he views as both exciting and concerning. He emphasizes the importance of mastering fundamental concepts in software engineering today, recommending understanding one programming language deeply, like Ruby or Go, and overlaying AI tools for productivity. He stresses the necessity of maintaining human taste and creativity, asserting that AI can generate artifacts that reflect human style and unique vision if steered correctly. Jankees discusses how AI is already capable of creating websites and interfaces, and he's experimenting with AI-driven design to produce distinct outputs. He advocates for responsible AI use and cautions about privacy and regulation, especially in Europe, where government policies could slow adoption but also protect user data.

The conversation shifts to how AI and programming can be integrated into education for children. Jankees shares his experience teaching programming to his kids, experimenting with simple game creation using AI tools like cloud code, which generates code for fun projects like changing a circle’s color. He emphasizes that children learn best by experimenting and exploring without fear of failure, and that this kind of hands-on, AI-assisted learning encourages creativity. He notes that his own children rapidly learn to type and code, often typing more efficiently than they write by hand, which highlights how digital tools can accelerate learning. Jankees advocates for giving kids access to modern technologies, enabling them to understand and build with AI at a young age, fostering curiosity and technical skills. He contrasts this approach with traditional education, emphasizing learning through doing and discovery.

Jankees talks about his involvement with Omarchy Kids, a project aimed at making technology accessible and fun for children. Currently working on a 'screen time' plugin that balances educational and recreational use of screens, he describes its two modes—strict and agreement-based—that allow parents to set limits while fostering trust and dialogue with their children. He shares his parenting philosophy against harsh punishments, favoring mutual understanding and discussions. The plugin features functions like showing time left, journaling for children to explain their screen time, and flexible weekly limits. Jankees envisions expanding the project to include features like website blocking, DNS management, and more, with the goal of teaching kids responsible, creative, and safe use of technology. He also discusses his broader goal of integrating AI and programming into early education, believing it prepares children for a future where AI will be central.

Throughout the episode, Jankees emphasizes the importance of community and open collaboration, mentioning his active participation in Omarchy’s ecosystem and sharing plugin developments publicly. He talks about the importance of security research, analyzing AI-generated code and exploits in isolated environments, and working with sophisticated security reports to identify vulnerabilities. He touches on the broader AI ecosystem, funding, and the rapid pace of development, noting that Europe faces additional challenges due to regulation and slower access to the latest AI models. He advocates for local AI models to address privacy concerns and mentions ongoing efforts in secure AI deployment, including his own experiments with local models and VM-based workflows. The episode concludes with enthusiasm for AI’s potential to reshape work and education, balanced with a cautious approach to regulation and ethical use.