I could not let August start without breaking down one of the biggest and strangely quiet tech releases of the year: ⁠chatgpt’s new agent⁠. If you have been hearing bits and pieces about agentic AI but are not totally sure what it actually means or why it matters, this episode will bring it into focus.

I walk through how generative, predictive, and agentic AI are different, how tools like Operator and Deep Research evolved into this new agent, and what connectors actually are and how they work. We also touch on real security considerations, like prompt injection, and how OpenAI is thinking about protecting users as these systems become more autonomous.

This is the patch:

• the difference between generative, predictive, and agentic AI
• how Operator and Deep Research led to the release of the agent
• what connectors do and why they change how AI can interact with your tools
• the security risks that come with more autonomous systems and how they are being addressed

At its core, this episode is about understanding what just changed in AI, without getting swept up in hype or panic.

We close with a lifestyle segment on my current work from home joys, including fresh sunflowers on my desk, the patchperfect summer smoothie the agent helped me create, and a few ⁠Olivia Dean⁠ tracks that are perfect for working sessions.

If you are curious about where AI is heading and want a grounded, practical way to understand what just launched, this episode is a great place to start. All views expressed are my own.