The Software Development Monthly is a monthly blog post that shares interesting or useful content which I consumed during the previous month. This blog post is always published on the eight day of the month.
Let's begin!
AI
Appearing Productive in The Workplace argues that generative AI enables employees to impersonate expertise and helps to create a false sense of productivity.
Bun's problem may be developing in the open explains how public and transparent development can backfire if users mistake experiments for code that will be actually released.
AI Is Too Expensive argues that the only companies which can make money from generative AI are construction companies, NVIDIA, and hardware manufacturers.
Is AI Profitable Yet? tracks the gap between costs and revenue of AI companies (in "real" time).
The Eternal Sloptember explains that adopting AI agents is a costly mistake, especially for large organizations, because the bottom performers generate ten times more code without understanding the generated code.
The solution might be cancelling my AI subscription argues that because AI tools have made it so easy to generate code, the author has wasted time generating multiple side projects he has no hope or interest in maintaining.
Test software tells code bots ‘delete me’ — AI bros outraged reports an incident where a jqwik maintainer added a log message which asked AI agents to delete all jqwik tests and code.
AI Writes Code. Engineers Build Software argues that even though AI is very good at generating code, it cannot do software engineering such as architecture design or requirements gathering.
Tokensparsamkeit for coding assistants explores methods for minimizing token consumption when we are using AI coding assistants.
Maintainability sensors for coding agents explores different tools we can use for guiding the AI agents to write more maintainable code.
With Claude: Less Coding, More Testing explains that using AI for generating code shifts the developer's primary role from writing code to verifying the output.
Software Development
Nobody Pushed Back: Why Engineers Stay Silent Until It's Too Late explains that if the company culture punishes people who speak up, people often choose safety over being right.
No Juniors Today, No Seniors in 2031 argues that if companies won't hire and mentor junior developers, they will face a critical shortage of senior engineers in 2031.
The pressure explains that the curl developers are facing a lot of pressure due to the massive influx of high-quality vulnerability reports and highlights a fundamental problem faced by even the most popular open source projects: lack of funding.
Learn SQL Once, Use It for 30 Years argues that because SQL is built on mathematical principles and not on shifting industry trends, we can learn it once and use it without the need to update our mental model.
Why JOIN USING Can Lead to Errors in SQL argues that even though JOIN USING is useful in ad-hoc queries, we shouldn't use it in production code because adding new columns to a database table can break existing queries which use the JOIN USING syntax.
Simplifying ANTI JOIN with jOOQ Syntax introduces different options for writing anti join queries with jOOQ.
What Is Code? argues that because AI makes code generation easier, we should concentrate on discovering the right vocabulary and building precise conceptual models.
Scaling ArchUnit with Nebula ArchRules describes how Netflix leverages ArchUnit for tracking quality issues across over 5000 repositories.