Vibe writing is what we should all be doing
AI is transforming writing education from rigid rules to a more fluid, experimental approach blending human and machine capabilities.
Large language models (LLMs) have fundamentally altered the landscape of written communication, creating both challenges and opportunities for educational leaders. As Rohit Krishnan argues in his provocative essay, "If AGI is the future, vibe coding is what we should all be doing"—a concept I'm extending to "vibe writing" for educational contexts. This shift requires instructional administrators to reconsider not just which tools we use, but how we conceptualize the writing process itself in an era where AI can generate sophisticated text at unprecedented scale and speed.
The integration of AI writing tools calls for a nuanced approach that neither dismisses these technologies as mere shortcuts nor embraces them uncritically. Simon Willison's characterization of LLMs as "overconfident pair programming assistants" provides a useful framework: these are collaborative tools that produce rapid output requiring human judgment to evaluate and refine. For educational leaders, this suggests a shift from teaching isolated writing skills toward cultivating critical evaluation and thoughtful curation—the essence of what I'm calling "vibe writing," inspired by Krishnan's analysis of vibe coding.
Educational institutions now have the opportunity to pioneer approaches that transform AI from a potential disruption into a catalyst for deeper learning. By drawing on concepts from uncreative writing (Goldsmith), Fluxus event scores (Bowen), and applying the principles of vibe coding (Krishnan) to writing instruction, educational leaders can develop frameworks that prepare students for a future where human-AI collaboration becomes increasingly central to professional communication. As Krishnan notes, "Every forward leaning tech company is already doing this"—our educational systems must follow suit.
AI with natural language
AI models function as collaborative communication partners that generate text requiring human judgment to evaluate and refine. This mirrors the Fluxus movement's event scores, where instructional texts were meant to be widely distributed and interpreted by others, creating a participatory creative process. For instructional leaders, this suggests a shift toward what Krishnan describes as having "part time AI trainer" in everyone's job description—developing students' abilities to direct, evaluate, and refine machine-generated content while maintaining their own critical perspective and creative authority.
Setting appropriate expectations is crucial when incorporating AI language tools into teaching and learning experiences. Following Goldsmith's principles of "uncreative writing," we must recognize that even when using appropriated or AI-generated text, numerous creative decisions remain: selection, arrangement, context, and purpose all require human judgment. Krishnan's observation that "Whether you want to or not, you're gonna change from being an Individual Contributor to a Manager" applies equally to writing—students must learn to manage AI-generated content rather than simply producing everything themselves.
For instructional leaders, AI tools offer opportunities to reimagine writing instruction as a process of curation and remixing. By teaching students to critically evaluate AI-generated language and understand what Goldsmith calls "the machine that makes the work," we can foster both technical proficiency and conceptual understanding. This approach requires a shift in assessment practices away from evaluating only original production toward valuing skilled direction, thoughtful selection, and ethical contextualization of machine-generated content—skills that will define success in future workplaces where, as Krishnan notes, we can "ask for pieces of code, copy paste it into your favourite IDE (Integrated Development Environment), and press run."
Embracing vibe writing
Andrej Karpathy's concept of "vibe coding"—which Krishnan describes as "asking the models to do things and accepting it immediately"—can be extended to "vibe writing" for educational contexts, creating a space where students can experiment with AI tools to generate and refine language through an exploratory attitude. As Krishnan notes, "naming is incredibly powerful, so this term has completely taken off. Vibe is of course the word of the century at this point." For educational leaders, this suggests curricular approaches that encourage students to develop an experimental relationship with AI writing tools, focusing on results and readability rather than process-based orthodoxy.
The boundaries between technical writing and creative expression blur in vibe writing, requiring instructional leaders to rethink traditional metrics of assessment. Rather than focusing solely on grammatical correctness or structure, educators might evaluate a student's ability to experiment, iterate, and integrate feedback. As Krishnan observes, "Any job which has sufficient data to train, which is almost every job, and has decent ways of telling if something is right or wrong, which is also a large enough number of those jobs, can't help but transform." This approach validates multiple paths to successful communication, recognizing that in professional contexts, the effectiveness of the final product often matters more than the specific process used to create it.
This experimental approach raises important questions about creativity and authorship that educational leaders must address. Krishnan's assertion that "This is the future that's being built right now" challenges instructional administrators to develop ethical frameworks that help students navigate questions of attribution, originality, and intellectual property in an age where, as Goldsmith suggests, appropriation becomes "just another tool in the writer's toolbox." These frameworks should prepare students for professional environments where human-AI collaboration is increasingly normalized and where, in Krishnan's words, "individual productivity will be a function of how much inference you can 'suck' from the models."
LLMs as cultural technologies
Henry Farrell and colleagues reframe large AI models as cultural and social technologies rather than mere computational tools, arguing that they function like writing or the printing press as catalysts for knowledge coordination and social transformation. This perspective encourages educational leaders to consider AI writing tools not just as technical resources but as cultural artifacts that reshape how knowledge is created, shared, and validated. This broader framing helps instructional administrators develop programs that integrate technology with humanities and social sciences rather than treating AI as relevant only to technical disciplines.
Understanding AI's cultural dimensions helps instructional leaders develop programs that bridge traditional academic silos. This interdisciplinary approach recalls the tension in Fluxus between what Bowen calls the "object/time antinomy" and the "choice/anti-subjectivism antinomy"—fundamental contradictions that energized rather than limited creative practice. Similarly, the tension between human creativity and machine generation can become a productive force rather than a limiting binary, fostering a more holistic view of communication. As Krishnan points out, "We're much wealthier than three decades ago, programmers do vastly different tasks helped along by automation and the congealed remnants of decisions made firm from the eras gone past."
By framing AI systems as part of broader cultural systems, educational leaders can prepare students to navigate a complex digital future where authorship, originality, and creativity are constantly being redefined. Krishnan's observation that "The world of work has already transformed" highlights the need for curricula that cultivate not only technical language skills but also critical cultural awareness, preparing students for workplaces where traditional roles and processes are rapidly evolving. This comprehensive view acknowledges that, as Krishnan notes, "Even if the AI train comes to a crashing halt for some reason, this trend will only slow and not stop."
Implications for educational practice
Instructional leaders should redesign curricula to incorporate AI-driven language projects that provide students with opportunities to experiment with appropriation and remixing. These projects might draw inspiration from Miranda July and Harrell Fletcher's "Learning to Love You More," which updated Fluxus event scores for the digital age through 70 assignments encouraging creative documentation of everyday life. Similarly, educators might develop prompts that guide students through collaborative human-AI creation, teaching what Krishnan calls the increasingly essential skill of being a "part time AI trainer" while fostering critical thinking and ethical consideration simultaneously.
Educational administrators should create institutional frameworks that establish clear guidelines for ethical AI use while encouraging exploration and innovation. This includes developing assessment rubrics that value thoughtful curation and critical evaluation alongside traditional writing skills, redesigning physical and digital learning spaces to facilitate human-AI collaboration, and establishing clear policies regarding attribution and intellectual property. These frameworks should recognize Krishnan's insight that "This is not just true of coding, it's true of an extraordinarily large percentage of white collar work" and prepare students accordingly for this rapidly evolving landscape.
Professional development programs should equip educators with both technical and conceptual understanding of AI writing tools. These programs should move beyond simple tool training to explore the cultural and philosophical dimensions of machine collaboration, helping teachers understand what Goldsmith calls the "dialectical image" created when human and machine intelligence converge. By preparing educators to guide students through these complex territories, instructional leaders can ensure that AI integration enhances rather than diminishes educational experiences, creating learning environments that prepare students for a future where, as one of Krishnan's commenters notes, "Jobs have still changed dramatically. They remained in some capacity, because there was value to human input at a higher level of abstraction, while the lower levels were delegated to automation."
References
Bowen, D. (2014). Fluxus. Encyclopedia of Aesthetics.
Chen, A. (2025, March 10). Vibe coding, some thoughts and predictions. Andrew Chen.
Farrell, H., Gopnik, A., Shalizi, C., & Evans, J. (2025, March 13). Large AI models are cultural and social technologies. Science -- Policy Forum.
Goldsmith, K. (2011). Uncreative Writing: Managing Language in the Digital Age. Columbia University Press.
July, M., & Fletcher, H. (2007). Learning to Love You More. Prestel.
Krishnan, R. (2025, March 26). If AGI is the future, vibe coding is what we should all be doing. Strange Loop Canon.
Willison, S. (2025, March 11). How I use LLMs to help me write code. Simon Willison's Newsletter.