# Human Expertise in the Age of AI Brute Force Balancing the bitter lesson of AI scalability with the essential role of human focus. - **Source:** [O'Reilly Radar](https://www.oreilly.com/radar) - **Published:** 2026-11-10 - **Topic:** [technology](https://brifing.si/c/technology) - **Original article:** - **On Brifing:** https://brifing.si/s/b1967514-951f-44c5-a7d0-990213fdb8eb ## The Bitter Lesson Reality Richard Sutton’s 2019 essay, The Bitter Lesson, argues that general methods scaling with raw computation consistently outperform human-designed expertise. The AI industry currently prioritizes this brute-force approach, often ignoring the necessity of human roles. Tim O'Reilly, while acknowledging this reality, advocates for creating economic mechanisms that keep humans central to future progress. ## Defining Human-Pilled Perspective Being human-pilled means believing that while AI models possess vast, generalized knowledge, they lack the specific focus provided by human consciousness. Humans act as the essential lens that sets AI potential into motion. Without human intent, AI remains a sea of possibility, unable to solve specific problems with meaningful, directed judgment. ## Context as a Lens O’Reilly Media’s internal experiments demonstrate that AI models provide different, more accurate results when filtered through specific expert knowledge. By using the Model Context Protocol, the company forces models to consult established expert texts rather than relying on general training data. This shift transforms AI from a generalist into a precise tool. ## The Power of Magic Words Context engineering involves using specific magic words, like GTD or the Eisenhower matrix, to trigger complex frameworks within LLMs. These terms carry immense, pre-loaded context that guides AI agents toward desired outcomes. Mastering these inputs is becoming the new frontier of programming, as it prevents agents from running in unintended, runaway directions. ## Organizational Productivity Challenges Many companies struggle to translate individual AI gains into organizational productivity. Trail of Bits CEO Dan Guido suggests that consistent configurations are vital to avoid wasting tokens on repetitive context-setting. Turning past failures into infrastructure, or scar tissue into systems, allows companies to compound AI benefits across their entire internal workforce effectively. ## Historical Precedent for AI The current AI landscape mirrors the early internet shift from filter-then-publish to publish-then-filter models. Google’s rise demonstrated that brute-force algorithmic sorting beats curated lists. However, this historical transition highlights that even with superior automation, the value of human experience remains in defining what matters, ensuring technology aligns with human needs. ## The Enshittification Risk Silicon Valley’s recent trend toward enshittification prioritizes corporate revenue over user experience, often sacrificing genuine utility for engagement metrics. This pattern threatens the AI economy by potentially alienating the very users who create the content models need. Maintaining human-centric design is essential to prevent these systems from becoming purely extractive tools. ## Monetization and Creator Rights The AI industry’s failure to share value with content creators contrasts sharply with YouTube’s Content ID success. By allowing rights holders to monetize rather than just block content, YouTube fostered a creator economy. AI companies currently focus on internal revenue, missing opportunities to build sustainable ecosystems where value circulates among all participants. ## Designing Economic Mechanisms Mechanism design, or reverse game theory, offers a framework for creating rules that produce desired social outcomes. Nobel laureate Al Roth applied this to kidney exchanges, proving that intentional design works. O’Reilly argues that AI platforms must adopt similar mechanisms to ensure the economy remains productive and inclusive for human workers. ## The Jockey and Horse A racehorse requires a jockey to compete effectively, serving as a metaphor for the human-AI relationship. AI provides the raw power and speed, while humans provide the direction and purpose. Without this partnership, the technology lacks the focus required to achieve meaningful results, regardless of how much computation is applied. ## Reliability and Industry Awe The AI industry is currently too in awe of its own progress, often ignoring the practical problems of implementation. Similar to past over-optimism regarding atomic energy, the current hype cycle risks overlooking the need for reliable, safe systems. Industry leaders must shift focus from pure progress to framing problems correctly for users. ## Future of AI Economics If AI platforms continue to displace humans without creating new economic affordances, they risk destroying their own customer base. Designing systems that allow users to monetize their own contributions is critical for long-term survival. The future of the AI economy depends on intentional mechanism design rather than accidental, market-driven outcomes alone. ## Sources covering this story - https://www.oreilly.com/radar/human-pilled/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._