Less Wrong, not more fast / How to fail fast?

A lightly edited ChatGPT summary of the session: The session discussed the concept of failing fast in order to be wrong for the shortest possible amount of time. It was emphasized that it is important not to focus on trying to succeed quickly but rather on failing quickly and learning from it. The participants discussed the timing at which they consider failing fast, even while gathering requirements. They highlighted the importance of learning early about the quality of the product through automated tests and identifying when it is not meeting expectations.

The conversation revolved around defining failure, distinguishing between doing the right thing and doing the thing right. The participants also shared their experiences of talking to customers/users and how it led to valuable insights. They mentioned the challenges of handling the overwhelming amount of information gathered during these conversations but acknowledged the opportunity to refine ideas and discard features that no longer seemed valuable.

The session touched upon the concept of building a "second brain" to gather and converge feedback effectively. A startup example was given where releasing code to customers every week helped in gathering feedback but presented difficulties in maintaining code quality.

The discussion also delved into the problem of human nature, with a tendency to resist throwing away code and a desire to leave an impact or legacy. Participants shared experiences of conversations with users that went off track and the importance of addressing simple questions and aligning expectations.

The session suggested getting developers to write unit tests and implementing automation tests and static code analysis to improve code quality. The idea of failing in small batches was highlighted, breaking features down into small increments of value.

There was also a discussion about the challenges of applying the fail fast approach in embedded systems, using examples from Space-X and Copenhagen rocket development.

Lastly, participants emphasized the need for psychological safety to embrace the fail fast mindset. They encouraged viewing experiments with negative results as successes and adopting the perspective that failures are opportunities for learning and growth.