Resilient Careers Rely on Bundling Tacit and Relational Tasks

The future of work is not about human versus machine. It is about the structural friction of messiness. Economist Luis Garicano argues that AI will not simply replace jobs. Instead, it will force a radical re-bundling of tasks. The jobs that survive and thrive are those where cognitive and relational tasks are so tightly intertwined that they cannot be separated without destroying the value of the work itself. For the reader, this reveals a non-obvious competitive advantage: the most resilient career paths are not those that resist AI, but those that leverage it to manage high-stakes, high-trust environments. The true risk is not automation. It is the potential for data monopolies to capture the rents of human expertise. Understanding this shift allows professionals to pivot toward roles that demand human judgment, providing a hedge against the commoditization of knowledge.

The Hidden Cost of Clean Efficiency

Conventional wisdom suggests that AI will replace jobs by automating discrete, repetitive tasks. Garicano’s analysis suggests a different dynamic: jobs are essentially bundles of tasks. A clean job is one where tasks can be neatly separated, making it highly susceptible to reinforcement learning. Conversely, messy jobs, like high-end sales or complex relationship management, possess synergies between cognitive and relational components. Attempting to automate the cognitive part of these bundles often causes the entire system to fail, because the knowledge required is tacit and deeply embedded in the interaction.

"The jobs that most clearly our listeners can identify are sales jobs, a relationship manager in a bank... there is a messy job because in order to perform the task, the cognitive part cannot just be peeled off and done by somebody else."

-- Luis Garicano

The implication is that clean jobs, such as transactional coding or basic contract drafting, are vulnerable to being unbundled and commoditized. The competitive advantage lies in remaining within the messy bundle, where the human element is a critical component of the value proposition.

The Looming Battle for Data Rents

While the overall pie of economic productivity may grow, the distribution of that value depends on who owns the data. Garicano warns that if knowledge is captured and distilled into proprietary AI models, the owners of that data will capture all the economic rent, potentially relegating the rest of the workforce to low-wage, disposable roles. This creates a systemic tension: the frontier models are essentially expropriating human expertise to train themselves, which could lead to a future of new feudalism where only a few incumbents control the means of cognitive production.

"Once they absorb the person owning the data will get all the rents and everybody else is disposable so the real land the new land are the data."

-- Luis Garicano

This shifts the focus from technological capability to political economy. The durability of one’s career may depend less on technical skill and more on the ability to operate in environments where data cannot be easily centralized or monopolized.

Why Slow Transitions Create Lasting Moats

History suggests that technological shifts are rarely smooth. The Second Industrial Revolution succeeded by standardizing parts and bypassing the artisan, but only after a long, painful transition. Garicano notes that we are currently in a period where organizations are trying to re-bundle tasks. The immediate, uncomfortable work of resisting the urge to clean up your workflow, by maintaining human-in-the-loop oversight and preserving relational nuances, creates a moat. Most teams will rush to automate, creating brittle systems. Those who patiently maintain the messy, human-centric parts of their process will find themselves with a durable advantage once the initial hype cycle settles and the limitations of clean automation become apparent.

Key Action Items

• Audit Your Task Bundle: Over the next quarter, categorize your daily tasks into clean (repeatable, data-driven) and messy (relational, high-trust, tacit). Identify which messy tasks you can double down on to increase your indispensability.
• Prioritize Tacit Knowledge: Invest in developing skills that are generated through interaction rather than static information. Focus on the how of your industry, the nuanced, relational approaches that are not written in manuals.
• Diversify Your Data Exposure: In the next 12 to 18 months, prioritize working with open-weight models or systems that allow your organization to retain ownership of its proprietary data. Avoid reliance on black box frontier models that may eventually commoditize your unique expertise.
• Advocate for Human-in-the-Loop Standards: In your professional sphere, push for internal policies that require human oversight for critical diagnoses or decisions. This creates a structural barrier to full automation, preserving the strong bundle of your role.
• Focus on Elastic Demand Sectors: When considering career pivots, look for sectors where demand is highly elastic, like the radiology example where lower costs lead to significantly higher usage. Avoid sectors where productivity gains lead to market saturation and wage stagnation.