Thanks for sharing! I wrote this after thinking about whether platform engineering actually something new, or just a rebrand and seeing this comment on reddit.
My personal hot take is that DevOps (in the truest sense of the word) is a dead end in the way Kelsey Hightower also sees Kubernetes as a dead end.
This isn't to say that DevOps isn't important to the computing world or that it hasn't done anything significant for the industry.
[But] I have yet to see this work well in practice beyond Series A startups where engineering staff count exceeds 20-ish people
I started in the this-is-just-a-fad-camp, where sysadmin became a devops engineer and then after people saying devops shouldn't be a title they became a platform engineer.
But, after talking to people ( and with my work being focused on interacting with legit teams and practices platform engineering ), I came out the other side realizing that pe teams and building a platform does solve a fundamental problem with pure DevOps. That "everyone does everything" utopia of the original DevOps vision didn't really scale to larger organizations, and you ended up with 'DevOps teams' anyhow. Not everybody can do everything. The human brain has limits and building tooling is a separate and useful thing.
But, after talking to people ( and with my work being focused on interacting with legit teams and practices platform engineering ), I came out the other side realizing that pe teams and building a platform does solve a fundamental problem with pure DevOps. That "everyone does everything" utopia of the original DevOps vision didn't really scale to larger organizations, and you ended up with 'DevOps teams' anyhow. Not everybody can do everything. The human brain has limits and building tooling is a separate and useful thing.