A public record of practice.
GVibeDev exists to show the work itself: what is being built, how it evolves and what becomes possible when strong intent is paired with better tools.
About GVibeDev
A personal workshop for games, tools, stories and experiments.
GVibeDev is the name I use for a personal workshop built around one simple idea: new tools can let people attempt things that once required skills, teams or resources they did not have access to.
It is not a software company, not an agency and not a brand dressed up as a team. The work begins with ideas, requirements and a willingness to learn whatever the next problem demands.
Background
Learning by building, not by pretending to already know.
GVibeDev grows out of multidisciplinary interests: writing, game design, systems-thinking, visual experimentation and the practical challenge of turning ambitious ideas into working artefacts.
I do not approach this work by presenting myself as a traditional software engineer. Much of the technical side has been learned inside real projects, through iteration, debugging, correction, comparison and the discipline of testing whether something actually works outside the excitement of the first draft.
That is why the portfolio matters. It is not a list of concepts. It is a record of attempts that became interfaces, tools, pages, game systems, documents and workflows.
Tools, method, responsibility
AI is leverage, not autopilot.
Language models can help write code, analyse systems, research a problem, reorganise a text, compare alternatives or accelerate prototyping. They can extend reach, but they do not inherit responsibility.
Intent, requirements, judgement, verification, acceptance criteria and final direction remain human responsibilities. The more powerful the tool becomes, the more important the method becomes. Fast generation without verification only produces mistakes faster.
That distinction matters here. The goal is not to remove the person from the process, but to give one person access to capabilities that previously demanded much larger amounts of time, specialised knowledge or external structure.
AI can amplify a process. It does not replace the need for clarity, discipline or responsibility.
This approach eventually became a project of its own: Manuale LLM, a practical attempt to turn long experimentation with language models into a repeatable method grounded in structure, responsibility and deliberate use.
Featured method project
Manuale LLM
A public writing / method project focused on how to work with language models using structure, prompting discipline, iterative diagnosis and human responsibility rather than vague automation myths.
A small shared workbench
Some corners of the workshop are shared.
My son sometimes joins me in experimenting with games, mechanics and wonderfully unreasonable ideas. His own public-facing projects live under the BiggityKotname, with their own tone and identity inside the wider site.
Working together is less about teaching a predefined formula and more about discovering what happens when curiosity is given tools.
The result is not a company-within-a-company, but a shared creative bench where related experiments can grow without losing their own personality.
Projects, questions, strange ideas.
For project discussions, repository issues, collaborations or general messages, the best next step is the contact page.