I let GPT-6 Astra run my business… it’s insane

In this deep-dive conversation, David Ondrej and his guest analyze the evolving landscape of AI agents, specifically focusing on the shift from human-prompted actions to autonomous agency. The discussion moves beyond surface-level AI use, highlighting how agents like GPT-6 Astra can be integrated into business workflows to handle everything from customer support and code reviews to complex logistics and infrastructure management.

The conversation examines the 'bottleneck' effect of human intervention and how delegating decision-making to AI agents can drastically increase productivity. By providing these agents with context, access to internal tools, and clear constraints, the hosts demonstrate how they are creating 'autonomous factories' where routine business processes can occur with minimal human oversight. They also share practical examples, such as how AI agents can autonomously identify and fix bugs, or optimize supply chains, while learning to respect human-defined safety rails.

Throughout the interview, the participants address the inherent tension between efficiency and control. They debate the risks of 'dangerous' AI behavior, the importance of maintaining human judgment, and the ethics of delegating critical decisions. The episode culminates in a look at the future of AI-driven commerce, where human intent acts as the core strategy, while AI handles the execution of thousands of micro-tasks, effectively scaling operations by orders of magnitude.