The One True Way to Build Software 

Great vid, love it - subscribed. Yet I feel one thing is missing at 10:15 ... Trust! Mutual trust, but especially top-down trust, that everyone is trying their best to achieve those four values.

Excellent video, Thank you Elizabeth. Just subscribed!

Thanks for the upload! Great points and perspective, well explained. I'm curious about your thoughts on how a company should handle promotions, and allow leaders to emerge and rotate. Sorry for the bad english

"The four ingredients that are absolutely essential" 10:14 1. "Culture that values and rewards collaboration, experimentation, and learning" - a.k.a. People should be encouraged to follow general heuristics that correlate with better individual rationality. 2. "Clear definition of what good looks like, or what's important for the context" - a.k.a. a well-defined reward function to be used by individuals to judge whether they are doing well or not (or whether others are doing well or not) 3. "Some way of giving everyone visibility into the current reality so that they can see the delta between the status quo and the ideal" - Figure out how to improve people's rationality of the project and the goal so they have a better model of the whole process to make improvements with 4. "Making time for teams to reflect and adapt... that is, hold restrospectives." - Provide opportunities for synchronization between different people's models of the process, the reward function (goal) specification, and rational consideration of what could be improved for the future. "Everything else is an implementation detail" Tbh, I think most of the principles above could also be implementation details, or at least they are derived principles. In my mind the most important thing is "being rational", and by that I mean you should embody just two principles: *Epistemic rationality*: systemically refining your internal model of the world around you to better predict it, a.k.a. "being accurate" *Instrumental rationality*: Using that model to achieve your values by figuring out what action to perform at any given moment that would maximize your prediction for your value(s) to be achieved (whatever those values may be). Everything else is a heuristic (a practice / thought pattern that is useful in a large number of cases)

I only develop apps as an Indie. I'm ruined as an employee for large businesses. Unemployable at this point. I like brewing my fussy coffee, coding in my pajamas at 2AM or 2PM or on a Sunday. To your point, "no right way", long as it ships =)

Elizabeth, have you ever worked with foreigners? Hugs from Brazil.