Scrum Made Zero Sense for My Young Startup. So I Designed an Alternative

Maybe I'm just unlucky, or it's a case of No True Scotsman, but I've yet to experience a good working Scrum environment that doesn't devolve into micromanagement, meeting hell and cargo-cult behaviors. I've seen it kill all the initiative, enthusiasm and curiosity in a developer team. Pre-Scrum, a developer might look at some code, think "hey, I'll spend a bit of time optimizing that, looks like it needs some tuning", maybe have a quick chat with their lead and/or teammates, and get on with it. Now it's "I'll have to make a card, which will probably be a spike, put it in the backlog, wait for it to be groomed, and maybe we'll get round to it in Sprint 12 this year". That becomes too much effort, so they don't bother. This is classic bureaucracy as a cudgel : make the form-filling and meetings exercise so burdensome ordinary people just give up, and the bureaucrats win because they thrive on the form-filling and meetings.

Scrum - as I've seen it practiced in the wild - seems to go against the entire spirit of agile and developer empowerment, and even calling it "agile" is pure gaslighting, but at the same time the arrival of Scrum is the signal that the company you work for is no longer the fun, challenging little startup you joined, but is now pulling on the big-boy enterprise pants. That may or may not be a good thing for the company as a whole, but as a developer it's the signal that you either should buckle up and prepare to be a more secure but unhappier cog in the new beaucratic machine, or move on to greener pastures. Because Scrum is just that: it's a tool for to corral your unruly startup weirdos into a growing "serious" enterprise organization, and to push out the ones who can't or won't comply. People just make the naive mistake of thinking it's about increasing productivity and efficiency.

Amplified learning is critical for a startup because you are searching for value in dark space. To find value you have to package up attempts and fire them into the dark space to see if anything hits.

Fire the value attempt when it is ready, not when some artificial sprint interval finishes!