Knowledge Work as Code
Everything is code: any knowledge work that happens on a computer can in theory be done with code, and it gets dramatically better when it is. The surprising corollary is about who benefits — the biggest winners of vibe coding are not programmers, they are everyone else, because the work that most improves when reframed as code (making a deck, a spreadsheet, a small tool) is the ordinary knowledge work non-programmers already do. Slides are the canonical example: on the surface a presentation has nothing to do with code, but done as code (HTML) it becomes beautiful, interactive, and far better — and it's the ideal on-ramp because there's no scary blank page (you already know what you're trying to make) and everyone has to make a deck at some point.
The practical arc: vibe-code your next deck, put your energy on the content instead of pushing boxes around, and over time you notice your own taste and workflow — which you can then freeze into [[skill-extraction | skills of your own]]. The thesis reframes democratization: value isn't just moving up off commoditized execution (Execution Commoditization, Imagination Constraint); it is moving out, to the non-technical people for whom "code" was previously a wall.
Claims
- Any knowledge work done on a computer can in theory be done with code, and it gets dramatically better when it is. principle — durable thesis; the "better" is load-bearing (the HTML deck is the demonstration), though the ceiling depends on the task and the model's fluency in the relevant code (HTML as Native Output).
- The biggest winners of vibe coding are non-programmers, not programmers. observation — the video's claim; a prediction about distribution of benefit, groundable over time, asserted here as the source's view rather than settled.
- Start with slides as a first vibe-coding project — no blank page, everyone needs a deck, and it moves your energy from box-pushing to content. best practice — context: non-technical people intimidated by vibe coding; slides are recommended precisely because the goal is already known and universally needed, lowering the on-ramp.
- Doing knowledge work as code lets you keep going until you notice your own taste and workflow, which you can then freeze into your own skills. observation — the bridge to Skill Extraction: repeated as-code work is how a personal workflow becomes visible enough to package.
Related
- Vibe Coding — the practice knowledge-work-as-code recommends: conversational build-and-iterate with a coding agent; this concept is the why it's for everyone thesis on top of it.
- Skill Extraction — where sustained as-code work leads: freeze your noticed taste/workflow into a skill of your own.
- Execution Commoditization — the adjacent value-migration claim: value leaves cheap execution. Knowledge-work-as-code is the democratization twin — value also moves out to non-programmers, not only up to task-selection.
- Imagination Constraint — once anyone can do knowledge work as code, the binding constraint becomes knowing what's worth making.
- Agent-Mediated Software — the delivery half: the code non-programmers produce is increasingly run and personalized through an agent layer.
- HTML as Native Output — a concrete reason as-code work pays off: the model's most fluent output format is also human-friendly.
- Distillate: Stop Making PowerPoints: Vibe-Coding HTML Slides as a Skill