Will A.I. Still Take Our Jobs?

The introduction of AI tools has drastically altered the process of generating professional work, exemplified by a team creating proposals that were highly polished in a short timeframe. Previously, lengthy research and development took weeks for an individual, but with AI assistance, a recent graduate organized brainstorming, used large language models to create research plans, delegated tasks to AI agents, and synthesized results, allowing a first draft to be completed in a week. This efficiency has led to competitors achieving comparable high-quality results, shifting the evaluation criteria for proposals away from assessing underlying intelligence or commitment toward assessing the final output.
This change introduces challenges for individuals, such as the original author, who struggle to prove personal value when expertise is rapidly commoditized. Furthermore, economists predict an increase in conflict resolution and authority-based decisions because tacit knowledge—the unwritten context of decision-making—is less accessible to AI systems. The proliferation of easily generated, high-quality content makes it difficult for managers to discern the source of true creativity or to gather the necessary tacit information regarding employee trust and ambition.
The implications extend to the nature of work itself, as AI changes production functions by enabling the creation of "90/10" quality, where reaching full excellence becomes difficult without specialized focus. This automation creates a split between "commodity tasks" where 'good enough' suffices and "star tasks" requiring true excellence, forcing workers and managers to redefine how value is measured.

* A team used AI to create a research plan by running brainstorming transcripts through a large language model.
* The team delegated parts of the research plan to AI agents and assigned fact-checkers.
* A first draft of a proposal, including visuals, was completed in one week after feedback.
* Frank's and Jill's teams also produced highly researched proposals with polished writing and images.
* Economists predict an increase in conflict resolution and authority-based decisions due to the ability to generate AI-based analysis.
* The difficulty lies in AI's inability to access tacit information, such as what a CEO truly thinks.
* AI application has been widespread, with office workers using it daily, transforming fields like coding, recruiting, and research.
* Job openings for software engineers increased in 2026, while customer-service jobs may disappear.
* AI creates a "90/10 production function," making it easy to achieve 90% quality but difficult to reach 100% excellence.
* Workers with "strong bundles" (where tasks are tightly integrated) gain value from AI in scenarios like "robots above" or "robots below."
* The scarcity remains in workers who can help organizations adapt to AI, focusing on understanding day-to-day work and workplace politics.

The narrative explores the tension between automated output and the human necessity for context, authority, and deep expertise. The central pattern observed is a systemic shift where productivity gains risk obscuring the underlying mechanisms of value creation, leading to increased friction in hierarchical structures. The "90/10 production function" acts as a powerful metaphor: automation excels at creating readily acceptable, high-quality surface results, but achieving true excellence requires non-automatable human judgment and iterative reflection that is actively bypassed by speed demands.
The dilemma for decision-makers arises because the ability to judge intrinsic qualities—intelligence, creativity, and ambition—is eroded when output quality becomes standardized. The Tacit knowledge gap is critical here; if AI handles the surface-level creation of documents, managers face an intensified need to discern *who* to trust, which shifts authority reliance from demonstrable competence to perceived relational history and network access.
The concept of "strong bundles" highlights a crucial structural implication: integrated roles demand longitudinal experience and relationship-building that is difficult to simulate through prompting. This implies that while AI can handle the execution (the commodity tasks), the organization’s capacity for true innovation and adaptation hinges on nurturing those long-term, relational worker attributes. The challenge for organizations is not simply adopting the tool, but redesigning reward systems and organizational structures to incentivize the development of these non-automatable, holistic skills—specifically fostering workers who possess both technical literacy and an understanding of workplace dynamics—as this group represents the necessary leverage for navigating the messy, emergent realities that AI cannot yet capture.
Bridge Questions:
What specific structural changes are necessary for organizations to reward 'star task' performance over mere compliance? How can managerial systems be designed to explicitly map and value the tacit knowledge embedded in strong bundles rather than relying solely on observable output metrics? If expertise is increasingly mediated by automated processes, what new forms of accountability must emerge to distinguish between automated efficacy and genuine human insight?