Am I Even Needed Anymore? GLM-5, Agentic Loops & AI Productivity Psychosis - EP99.34
Thursday, 12 February 2026 · 2 min read · Listen to the episode ↗
The hosts explore the phenomenon of "AI productivity psychosis," highlighting the pressures of constant productivity fueled by AI advancements. They discuss GLM5, a promising yet cautiously evaluated AI model, and stress the importance of enhancing agentic loops for efficient task management, while acknowledging challenges in integrating AI into workflows. The emotional toll of these shifts raises existential questions about human relevance in technology-driven roles, amid concerns of cognitive overload and worker fatigue.
The hosts discuss the impact of AI on productivity, coining the term "AI productivity psychosis" to describe the pressure to constantly be productive. They focus on GLM5, an open-source AI model from a Chinese lab, expressing skepticism about its serious applications while acknowledging its potential to shift paradigms. Initial feedback on GLM5 is positive, with it performing comparably to other models in basic tasks, though caution is advised regarding potential rough edges with extensive use.
The conversation highlights the effectiveness of AI for routine tasks, particularly in agentic tasks like drafting emails and managing workflows. They note a competitive landscape in AI development, emphasizing the importance of enhancing agentic loops for task execution. Feedback on agentic-tuned models reveals challenges in integrating them into workflows requiring conversational exchanges, as they can disrupt normal interactions.
The hosts reflect on the human element in coding, stressing the importance of testing and validation despite using tools like git. They share strategies for managing cognitive overload and the necessity of user validation. The discussion also touches on the efficiency of AI, with one speaker expressing frustration when AI underperforms, drawing parallels to high expectations set by technologies like Tesla's self-driving features.
Concerns about cognitive overload and the potential for future technology to alleviate it are raised, alongside discussions about the evolving nature of work and the role of task management tools. The hosts express worries about increased multitasking and productivity expectations leading to worker fatigue and anxiety.
The emotional impact of AI on work is a significant theme, with one host voicing existential doubts about their relevance in a changing environment. They mention recent departures from major AI labs due to safety issues and clarify misconceptions about AI agents, asserting that interactions are driven by user prompts rather than independent goals. The conversation concludes with reflections on the implications of AI capabilities and a sense of disillusionment as individuals pursue personal passions over their roles in technology.
This summary was generated from the episode transcript and can contain mistakes.