"Vibe Work": A New Paradigm for Knowledge Work?

Author note: The following article is my attempt to generalise the points I made in this earlier article - I think what is happening in coding is something of a template that we may see play out in other domains.

This is a rather assertive take with very little experience to back it up. Only what I see in front of me and a few months in the rearview mirror. But that is, at this moment, all any of us have. The ground is moving very fast, which is exhilerating, but also destablising for many (including me), and a generator of anxiety. Ironically, or at least, interestingly, I also find that this means there is very little time to ponder anymore. I feel instead that I need to quiclky observe, summise, own my assumptions and execute in quick iterations. Anything else feels too slow. As long as I can try and keep a clear head, adjust and iterate, I might have a chance to ride this storm.

AI introduces a fundamental shift in how we interact with professional tools and processes, moving from deterministic workflows to a more fluid and adaptive paradigm. Originally termed Vibe Coding by Andrej Karpathy in February of this year to describe a new approach to software development, the concept extends well beyond programming. This broader idea—Vibe Work—captures a similar shift in knowledge work across multiple domains.

The Nature of Vibe Work

At its core, Vibe Work is defined by:

• A persistent gap between human intent and AI-generated outcomes.
• A shift from deterministic control to probabilistic collaboration.
• A willingness to engage with outputs that diverge from the original goal but may still hold value.

The ability to work well in this environment requires iterative interaction, collaboration with AI, a willingness to accept unexpected outcomes, and the capacity to evaluate and shape results as they evolve—rather than expecting predictable, finished outputs.

Three Strata of Adaptation

As AI enters professional domains, three broad adaptation patterns are emerging:

1. Novice Innovators
These are individuals with little or no prior domain expertise who are now producing surprisingly advanced outputs with AI. A liberal arts graduate may build functional apps or a small business owner might design marketing videos. With no legacy habits or rigid expectations to overcome, they often embrace Vibe Work intuitively—adapting quickly, iterating freely, and discovering new possibilities without hesitation.

2. Adjacent Professionals
These users are familiar with a domain but have traditionally relied on others to implement their ideas—such as managers, strategists, or designers without coding or analytical skills. AI allows them to close the gap between idea and execution. A strategist can now test campaigns, a product manager can mock up interfaces, or a consultant can run detailed analyses directly. They balance adaptability with enough domain understanding to meaningfully guide the AI, often discovering newfound autonomy in the process.

3. Traditional Practitioners
This group consists of highly trained experts. They are the most technically equipped to evaluate AI outputs, identifying subtle errors or inconsistencies that others may miss. But that same clarity can make adaptation more difficult. Seeing where and how AI fails may increase skepticism, making it harder to surrender to open-ended collaboration or accept "good enough" results. While they bring irreplaceable depth, they may struggle most with the looseness that Vibe Work requires.

Skills for Vibe Work

The core competencies for Vibe Work differ in emphasis from traditional skills. It's not that traditional practices lack these traits, but they are more central—and more frequently exercised—in Vibe Work. These include:

• Curiosity: A disposition toward exploration and questioning rather than clinging to certainty.
• Adaptability: The ability to shift direction in response to unexpected outputs.
• Iterative Refinement: Engaging AI not as a static tool, but interating through extended back-and-forth refinement.
• Tolerance for Ambiguity: Recognizing value in incomplete, strange, or surprising results.

These skills apply across all three groups, but they manifest differently. Novices may adopt them quickly by default, because they lack fixed expectations. Adjacent professionals bring contextual knowledge and often thrive as they gain direct creative control. Traditional practitioners may find the skills less intuitive to adopt—not because they lack the capacity, but because they actually 'see' where AI 'got it wrong'. For them, the core tension of Vibe Work is most acute.

Where to From Here?

It’s worth closely observing what’s happening in coding, as it’s the most high-profile expert domain currently being transformed by AI. The rapid evolution of tools and practices, and their effect, in software development could offer valuable insights for other areas of knowledge work.