IMHO, too many people chase the CTO title but few really want the actual job.
So true. CTO being so far upstream, mistakes made here through disconnection between the rubber and the road have compound effects downstream. It's the sort of role where bad decisions have huge, company-ending ramifications.
"I talked to my buddy at [FAANG] and he said microservices are the way to go. His team achieved awesome velocity by switching to microservices. Let's do it guys!".
Amazon had huge dependency problem, both technical and organizational. Technically, the effort to work in their monolith scaled superlinear with respect to the CL size due to coordination and review with other teams. Organizationally, teams were dependent on an ever growing set of centralized, global processes that further slowed their ability to make progress.
Jeff (and S-team) saw these issues and worked to solve them through a series of experiments. Specifically, Jeff tasked his then CIO with finding a solution. This CIO worked throughout the organization to gather on-the-ground feedback around what was and wasn’t working; built a model for how Amazon _should_ work based upon that sense making; and then tested, iterated, and refined that model into what we’ve come to know as Single-Threaded Leaders and Two-Pizza teams, etc., today.
In parallel, a couple Amazon business units were experimenting with exposing their data via textual (XML) APIs like the Amazon Associates API which went on to become the Amazon Product Listings API. These early APIs were sort of the POCs that allowed Amazon to see early validation around the concept of web services.
Synthesizing all these different streams of information, ideas, and actions lead to Jeff’s “edict” around building standalone web services (which eventually led to what we know as AWS today).
In short, Jeff is the exact antithesis of GP’s straw CTO. He speaks from data, facts, and information gathered from a myriad of internal and external sources, not merely anecdotes of success from a CEO buddy. :)
Source: Working Backwards has a lot of information on Jeff (and Amazon’s) decision making.
While a CTO needs to have grown up in the technical trenches to have credibility, a CTO should not be making any technical decisions anymore. If they are, that's a red flag for a CTO who can't let go and keeps micromanaging.
A common pitfall among startups is to give the CTO title to whichever cofounder is leading the development when the team formalizes titles. Some times this person doesn't even have previous engineering manager (EM) experience at all, but as the most technically oriented person of the time they receive the CTO role by default.
Some times these people can grow into the necessary delegation, recruiting, hiring, performance review, people management, communication, and meeting skills necessary to be a CTO. Other times, they cling to what they know (coding) and turn into a micromanaging CTO who won't cede control of the things they want to do (code) while avoiding the things they have to do (managing).
This is one of many reasons why I advocate for avoiding C-level titles as long as possible at a startup. You can always promote someone up to CTO unceremoniously when the company actually needs defined C-level executives and they've been excelling at the role already, but it's much more problematic to ask someone to give up the CTO title and step aside to bring in an experienced CTO when necessary. Nobody likes being demoted, even if it's only a formality because they weren't actually doing the full management role. Any title demotion at startups is likely to lead to conflict and departures.
This article reads like something a startup CTO would do or expected to do. Frankly companies below 100 shouldn't have this role. Someone leading a 10 person tech team isn't really a CTO because the work and decisions involved are totally different from someone leading a larger tech organisation. They should just mark such roles as what it is: a senior architect.