Week #1
Started "Systems Performance"
Went through the first 3 chapters so far, but it feels like it is going to be an amazing read. What really stood out is the proposal of a systematic approach to the topic of performance and its assessment. Here are some sidenotes I've made so far.
"Level of Appropriateness"
Return-on-investment matters when it comes to the question "how far should I go with performance improvements".
"Modeling" and "Capacity planning"
A solid intro to capacity planning. This is the exact read that I will get back to once I need to do any estimation of service capacity ever again.
"Methodologies" chapter
Thorough overview of methods you could apply to measure, assess, understand and improve performance of various systems.
Good methodologies that I've heard before and described in a detail in the book are USE and RED methodologies, but I also liked the recap of all the basic methods which are useful for the performance analysis.
Watched "The Identity Crisis of Software Engineers"
Some smart guys discuss vibe-coding and AI assisted programming in general and its consequences. I liked the vibe (no pun intended) of the conversation and how balanced it is between Borislav (somewhat AI-pessimistic) and Rares (somewhat optimistic) points of view.
I do not fully agree, but can relate to the point that the software industry as a whole is going through some sort of 'identity crisis'. If before the 'AI era' software engineers' value was in ability to comprehand whatever is the state of the system in question, how to operate it without causing downtime and how to build theory1 around it. Now I feel like that speed of building anything workable may have become the most valuable skill.
One theme that was really to the point was dynamics of learning while programming using AI extensively. If you don't type out and just tab-tab-tab or prompt your way through new programming tasks - there isn't much one would learn from it without trying to actually think through and evaluate (as it running it, writing or constructing tests for such code "by hand") the machine-generated code.
Iakov's pick of the week
Why this website even exists?
I decided to write little reviews on things I read and learned this week. I did this more or less regularly for the first half of the last year (in Russian), so I decided to resurrect this format. As a bonus point now, Iakov will be leaving a song of his choice each week.
Footnotes:
Picked this term from the Peter Naur's essay Programming as Theory Building