You Must EXPERIMENT To Find Your Best Design 

You can never really fully comprehend how something is going to work until you bring it into existence and smack it around a bit.

You two talking is great. Memories, tactics and family life still there 😂🎉

First time in live i went slow in playback speed! Such passion.

You must create version 2.0 to fix what you realized that you did wrong in version 1.0

I've always thought of Engineering as Creativity + Applied Science

Love listening to Kevlin. And dear god, the amount of headaches we could save by simply taking an honest deep look at the real environment we are delivering to and simply adjusting for reality, even if that means we have to walk back on the "ideal designs" we wanted (but work real bad)... My current one being something that designed to "look like this" without knowing if a) the platform could deliver that and b) if so, AT WHAT COST? Much as i hate Apple, the one thing i like is that, if you don't see a simple and direct way to do something in the UI SDK even though it's doable, that's them silently advising you AGAINST doing it, because 99% of the time you're just trying to be "different for the sake of being different" and shooting yourself straight in the foot...

I had to call my internet provider for more bandwidth to handle two all-time greats in one video.

Idealogues will never understand this.

I love your reaction to the idea of Kevlin "meditating upon the design"... I wonder if you've gone to a similar visual place as I did... Kevlin, sat in the lotus position on a desk for 3 days with an "Illuminati triangle" slowly materialising above his head. :)

One of my previous employers tried to factor "experiments" like these by calling them something else in their "Scrum" process (I already forgot what they called them) and then we've had to estimate these in the team. We really sat there in estimation meetings to talk about how long anyone should invest into looking up solutions to specific problems. So if you picked one of these tasks, you'd have for example 24 hours (equaling three work days with 8 hours each) to look into this issue. If you figured out a solution sooner, you could document these findings and then create a task for the actual implementation. If not, you should then show your progress to the rest of the team and discuss how to continue. I think it is needless to say that nothing got done in that company and I left out of frustration because I got less done there in four full weeks than on one evening in my free time on my own projects (I wish I was exaggerating).

Nobody is picking up on the fact the architect is actually coding. This was the norm late 2000s into early 2010s, and unfortunately we've gone back to architects being useless and causing more problems than they solve by not coding.