Serendipitously, a series of fortuitous circumstances put me in touch with the CTO of a company in California who hired me as his mentor and coach. We’ve been at it for a few weeks and I am very happy to see him work through major obstacles with my help. Hearing him explain earlier this week how our chats have influenced him and how he is happy to share these experiences with other people in his life was incredibly validating.
❦❦❦
The main long-form reading from last month is this one:
- David Brooks - How America Got Mean
- The author argues convincingly that dropping moral education from the school curricula was the mistake that explains a lot of today’s dysfunction in US culture and politics. The article looks a bit long at first but actually covers a lot of ground and nuance, so it’s actually quite enlightening and worth a read. It also contains a powerful call to action. I found that part resonates with some of the thoughts I developed during the last year.
❦❦❦
Some bits and pieces about engineering management:
In The carefulness knob, Lorin Hochstein presents a short and powerful story about a powerful academic result, the Efficiency-Thoroughness Trade-Off (also called ETTO) principle: the cost to do something well increases super-linearly with how well we desire to do it.
In his talk Egoless engineering, Dan McKinley explains how to structure organizations to avoid excessive diffusion of responsibility. He shares his experimental distrust of component-based team boundaries and of project-based team boundaries. Key sentence in the presentation:
Strict division of labor feels obvious, but it is wrong, and also sucks.
He then also shares tips on how to increase collaboration and engagement and how to develop intentional values for teams.
In a related note, prolific author Ludic, whom I believe is called Nikhil Suresh, also shared Be Angry. In this thoughtful piece (albeit slightly inflammatory), he highlights how we should not accept mediocrity and how we should recognize and be optimistic about the power of coordinated endeavour instead.
In another related note, a person whose name is Steven but with an otherwise guarded identity shared, in Care Doesn’t Scale, a poignant argument that certain social needs cannot be well addressed by scalable solutions, and that we should find sustainable incentives to ensure they continue to be addressed even as technology continues to eat the world. Choice quote:
For care, though, it doesn’t get bigger and better. If your goal is to educate the world, you can look for ways to educate thousands or millions. If you want to inspire the world, the billions await. But if your goal is to care for the world, and in a given moment you’re deeply caring for one person, you’re doing the best it’s possible to do.
This article generated a healthy conversation on Hacker News, where a specific sub-thread caught my attention, with another choice quote:
[…] when the author says care “doesn’t scale”, they obviously mean “you need a one-to-one ratio of caretakers”, which I fully agree with. But what they’re also accidentally doing in the process is explaining why creating bigger teams with bigger hierarchies and structures does not appear to increase the efficiency of care.
And there is yet another sub-thread below that that reflects on incentive mismatches (often, between executives and rank-and-file) and how middle management is often in the uncomfortable position to have to reconcile the two.
❦❦❦
The only two things about generative AI you need to read this month are:
- Addy Osmami - The 70% problem: Hard truths about AI-assisted coding
- Where the author highlights that the generative AI tooling only solves 70% of real problems and that the remaining 30% require experience and knowledge. As more and more non-experienced people get access to this tooling, we will see a glut of quickly prototyped technology that only works 70% of the time, and causes major problems the rest of the time, and it will be hard to troubleshoot. His prediction is that as more people realize that, there will be a renaissance of software development as a craft, where people who deeply care about quality will become popular again.
- Anshul Ramachandran - The Most Dangerous Thing An AI Startup Can Do Is Build For Other AI Startups
Where the author highlights that enterprise software has design requirements that are often unknown to people who build startups fresh out of school, or who only have experience with online B2C services, so that all the technology produced by these people (especially related to generative AI) are deeply inadequate to the real world. He dives into several of these areas with non-trivial design requirements including security, compliance, analytics, latency and scale.
Reading this made me better appreciate how my engineering degree at school did, in fact, provide me very early on with a solid intuition about those things and a high level of expectations for the products I build, and I also realized that teams I am part of or that I would form would have an edge in the market for our ability to deliver on these aspects.
❦❦❦
References:
- David Brooks - How America Got Mean.
- Lorin Hochstein - The carefulness knob.
- Erik Hollnagel - Efficiency-Thoroughness Trade-Off.
- Dan McKinley - Egoless engineering.
- Ludic - Be Angry.
- Steven - Care Doesn’t Scale.
- Addy Osmami - The 70% problem: Hard truths about AI-assisted coding.
- Anshul Ramachandran - The Most Dangerous Thing An AI Startup Can Do Is Build For Other AI Startups.