We keep hearing that AI will “take work off our plate.” The truth is subtler. AI can take tasks off our plate, but only if we keep ownership on our shoulders, and build the guardrails that make delegation safe. ------------- Context: The Two Extremes We Keep Falling Into ------------- When teams start experimenting with AI that can take actions, sending messages, updating records, scheduling tasks, changing settings, the first reaction is often excitement. The second reaction is usually a control reflex. We either lock the system down so tightly that it cannot help, or we loosen it too much and hope it behaves. Both extremes are understandable. Over-restriction feels responsible because it minimizes risk. Over-delegation feels productive because it maximizes speed. But both approaches tend to break trust. Over-restriction produces disappointment and abandonment, because people cannot feel value. Over-delegation produces incidents, because the system will eventually act confidently in the wrong direction. The heart of the issue is that many organizations treat AI delegation as a binary choice. Either the AI is “allowed” or it is “not allowed.” That framing misses how delegation works in real life. We do not give a new colleague full authority on day one, and we do not keep them in permanent trainee mode either. We expand autonomy gradually, with clear boundaries and feedback. Agentic AI forces us to become designers of autonomy. Not in a technical sense, but in an operational sense. We are shaping what happens when the system is uncertain, when the context is incomplete, and when the consequences are real. ------------- Insight 1: Autonomy Needs Shape, Not Just Permission ------------- When we give AI the ability to act, the default state becomes momentum. The system will follow instructions, connect dots, and execute steps faster than we can think through edge cases. That is the point, and it is also the risk. So the key question is not “Should AI have autonomy?” It is “What shape should autonomy take?” Autonomy without shape is chaos disguised as productivity.

This week, I break down the Claude Opus 4.6 release from Anthropic and shows off some early test results, explains why so many are freaking out about Google's Project Genie, discusses OpenAI's latest Codex app release, and more. Enjoy!

Every entrepreneur has two voices running in the background. One voice wants progress. It pushes you to decide. To move before you feel ready. To take the next step instead of waiting for certainty. The other voice sounds reasonable. It tells you to wait. To gather more information. To avoid making the “wrong” move. Here’s the problem: That second voice doesn’t feel like fear and instead it actually feels like strategy. But over time, it quietly turns into regret. Most people don’t stall because they lack ambition. They stall because indecision feels safer than ownership. Progress is rarely blocked by effort. It’s blocked by hesitation. So here’s the question I want you to sit with today: What decision are you avoiding because you don’t want to be responsible for the outcome? Write it down. Then ask one more question: What is this costing me every week I don’t decide? Lost momentum? Lost confidence? Lost time? Lost learnings? Action for today : Pick one decision you’ve been circling. Decide it. Give yourself seven days to observe and adjust. You don’t need perfect certainty. You need movement. I promise you... this is the way.