The rise of 'vibe working' (www.businessinsider.com)

“Vibe working” has emerged as a cultural shorthand for using generative AI to make white‑collar work feel improvised and effortless. Big tech execs and startups are leaning into the language: Microsoft rolled out “vibe working” features that use agentic tools in Excel and Word to “speak Excel” and “vibe write,” Mea and Sora are surfacing AI‑generated video feeds and “vibe creators,” and executives tout AI for rapid prototyping and marketing. In practice, the trend reflects a shift from manual craft to AI‑assisted workflows—developers increasingly review AI‑generated code rather than author everything themselves, a dynamic dubbed “vibe coding,” and employers prioritize AI familiarity (a Microsoft report found 71% of leaders would favor less experienced hires with AI skills). For the AI/ML community, the rise of vibework underscores key technical and operational needs: robust human‑in‑the‑loop interfaces, reliable model behavior, reproducibility of prompts and outcomes, and explainability so outputs don’t become “workslop” (slick but shallow artifacts). There’s a large training gap—most workers lack formal AI upskilling—so best practices are being invented bottom‑up, creating inconsistent processes and measurement challenges. The implication is clear: to make “vibing” productive rather than performative, teams need improved tooling for prompt engineering, evaluation metrics, guardrails for quality and bias, and organizational investment in training and workflows that treat AI augmentation as skilled labor, not magic.