Brendan Foody on Teaching AI and the Future of Knowledge Work - Conversations with Tyler Recap

Podcast: Conversations with Tyler

Published: 2026-01-07

Duration: 1 hr 1 min

Summary

Brendan Foody, CEO of Mercor, discusses the innovative approach of employing expert poets and professionals to enhance AI models in creative and knowledge-driven tasks. The conversation explores the challenges of measuring AI performance and the future of knowledge work as AI continues to evolve.

What Happened

In this episode, Tyler sits down with Brendan Foody, the CEO and co-founder of Mercor, an AI company that has made waves since its inception in early 2023. Brendan, noted for being the youngest unicorn founder, shares insights into Mercor's unique strategy of hiring top-tier experts, such as poets, to guide AI models in creative writing tasks. He explains that the $150 per hour rate for poets stems from the need for quality evaluation and instruction, allowing these experts to craft rubrics and assess AI-generated poetry to better align with user expectations.

The discussion shifts to the selection of experts for AI evaluation, highlighting the importance of combining academic knowledge with real-world experience. Brendan emphasizes that the disconnect in AI research often lies in its focus on academic evaluations that don't reflect the practical outcomes necessary for industries like law and medicine. By surveying experts and analyzing how they spend their time, Mercor aims to quantify the economic impact of AI in various fields, showcasing significant improvements in model performance over time, particularly with recent advancements like GPT-5.

Brendan also addresses the limitations of AI models, noting that while they can excel in specific tasks, they struggle with integrating multiple skills and managing long-term projects. He hints at upcoming developments in Mercor's Apex project designed to tackle these challenges, suggesting that the next wave of AI innovations will enhance its capabilities in knowledge work. Overall, the conversation paints a picture of a rapidly evolving AI landscape where human expertise remains indispensable.

Key Insights

• Mercor employs experts like poets to enhance the performance and evaluation of AI models in creative tasks.
• There's a significant gap between academic evaluations of AI and the practical applications needed in industries like law and medicine.
• AI models show substantial improvement year over year, but their integration and long-term task management capabilities still lag behind human abilities.
• The future of AI in knowledge work will likely focus on bridging the gap between task-specific performance and comprehensive skill integration.