Following closely after my recent first edition, here’s my second.

I’ve recently been given LLM access at work, though fortunately with no particular minimum usage decree (yet). After initial trepidation about LLMs, I can say that I’m starting to better understand their appeal and some of their limitations. In particular, it’s making working with Ruby on Rails much more pleasant and productive, which I definitely appreciate. Caution is necessary: you must check LLMs’ work, as they are not nearly as trustworthy a partner as a compiler is. However, they indeed are a great additional tool in engineers’ toolbelts. I look forward to digging in more at work.

The articles follow. Please keep in mind that I am not presenting them because I necessarily agree with everything in them. They simply represent information that I felt was interesting or helpful at the time that I noted the article.

  1. Web development is fun again
  2. Why AI is pushing developers toward typed languages
  3. AI Coding Assistants Are Getting Worse: Newer models are more prone to silent but deadly failure modes
  4. Why We’ve Tried to Replace Developers Every Decade Since 1969
  5. Agent Psychosis: Are We Going Insane?
  6. Designing AI-resistant technical evaluations
  7. How AI Impacts Skill Formation
    • Summary: “Novice workers who rely heavily on AI to complete unfamiliar tasks may compromise their own skill acquisition in the process. We find that AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average. Participants who fully delegated coding tasks showed some productivity improvements, but at the cost of learning the library.”
  8. An ode to “Slowly” handcrafted code
  9. Competence as Tragedy + Reddit
  10. ChatGPT Is a Blurry JPEG of the Web: OpenAI’s chatbot offers paraphrases, whereas Google offers quotes. Which do we prefer? (Archive)
  11. AI Makes the Easy Part Easier and the Hard Part Harder + Reddit
  12. AI Coding Killed My Flow State + Reddit
  13. An AI Agent Published a Hit Piece on Me
  14. Twitter post about the actual AI situation + Reddit
  15. Amazon service was taken down by AI coding bot (December outage) (Article copy in Reddit comment)
  16. Agentic Debt
  17. Yes, and…
  18. AI was supposed to save coders time. It may be doing the opposite: Studies find AI helps developers release more software—while logging longer hours and fixing problems after the code goes live + HN + Reddit
  19. Which programming languages are most token-efficient? + HN
  20. A sufficiently detailed spec is code + HN + Reddit

<
Previous Post
On The Brevity of Ruby Code
>
Blog Archive
Archive of all previous blog posts