The way we build and ship code is changing. We’re all leaning more on AI in our day-to-day, from delegating mundane tasks to helping us think through problems. What happens, though, if we defer all thinking to LLMs? We explore that in our latest article on Cognitive Surrender.

Luca sat down with David to talk about his journey building EventCatalog, hard-learned lessons on building event-driven architecture in production, and how MCP is changing the way we understand architecture.

And to wrap up, we’ll be in Denver and Copenhagen later this month at conferences. We’re also hosting a private gin & whiskey tasting event in Copenhagen. Meet the founders (Jacob & Victor) and Luca. Register here (discount code below).

Hope to see some of you this month! 👋

“Cognitive Surrender” is a new phrase coined in a recent paper by Steven Shaw and Gideon Nave at the Wharton School (UPenn) called “Thinking, Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender”.

This paper extends the original thinking model (System 1 fast and instinctive & System 2 slow and deliberative) from the book “Thinking, Fast and Slow” and suggests that individuals now have a third system, System 3 (AI). The authors ran three experiments across 1,372 participants who had access to an AI assistant while solving reasoning problems. They showed results where participants consulted AI on over half of trials and adopted its answers without critical assessment. Even when the AI was wrong, they followed it on roughly four out of five trials when they consulted it. This creates a problem where individuals who “delegate tasks” are more tempted to “delegate decisions”, thus not developing their cognitive and reasoning skills.

Cognitive surrender is defined as the tendency to “defer judgment, effort, and responsibility” to AI output, particularly when it’s delivered fluently and confidently. In other words, the more ‘elegant and fancy’ the output, the more likely people are to take it at face value.

This pattern shows up among software engineers and architects too. With more AI tooling and increasing requirements from organisations, engineers and architects have the temptation to “delegate” some of their work to AI tools (e.g., code generation or writing spec) in order to meet demand. However, this “delegation” can easily fall into cognitive surrender if users are not carefully reviewing the output of their tools.

To draw the distinction, there are two behaviours:

David’s been following a similar journey to us, building a bootstrapped business for technical leads, engineers, and architects. In this episode, David talks about his nearly decade-long experience building event-driven architecture and how context is becoming the differentiator for agentic workflows.

Some upcoming events to check out for architects, team leads, and engineers. We’ll be at both of these events, so look out for us!

We’re excited to host a special gin and whiskey tasting in Copenhagen with Luca and the founders. Register here. Tickets are €20 (the actual cost to us is ~€60 per person), and we’re offering a 50% discount code to the first 10 registrants.

Use STAYCHILL26.

Stay chill,