2026-02-22
AI fatigue affects 71% of workers who say new tools appear faster than they can learn them. Learn what causes AI exhaustion, why experts suffer most, and practical strategies to reclaim your focus.
2026-01-28 | #ontology
I have developed my own YouTube video transcript RAG. I wanted to read summarized transcripts as blog posts, but when I was at the stage that the blog posts appeared in my blog right after a video that I had in my subscriptions was published I thought that I also would like to have all the content accessible for AI so I could ask an LLM about topics of the videos my system had consumed and an LLM would answer with all the references back to the videos where a piece of information was found.
2026-01-19 | #ai #ai engineering #guardrails #pydantic-ai #software development #spec driven development #symfony-ai #vibe-code #workflow
A Makefile orchestrated a couple of Python scripts. For a few days, my setup felt almost boring — in a good way. I copied a YouTube video ID, pasted it into a Makefile, ran a command, and with a prompt to analyze the transcript, an AI produced Markdown and wrote it to a file. I copied the file to a GitHub repository, pushed it, and a static site generator turned it into something I actually enjoyed reading: HTML in the browser, and an RSS feed I could follow like a personal newspaper.
2026-01-17 | #ai #ai engineering #application #application layer #workflow
I have a busy family life and no time to watch hours of YouTube tutorials. Also, I find the information density very sparse and extracting the information very time-consuming. My “Neo from the Matrix” fantasy has always been to have knowledge injected directly into my brain—skipping the medium and going straight to the data. So my wish formulated that I would like to have the information extracted, so I can expose my brain to the compacted knowledge and speed through all the knowledge I would like my brain to have consumed.
2025-11-03 | #architecture #hexagonal architecture #php #tdd #testing
A good test should tell me if a feature is broken. But I have worked with many tests that failed only because I moved a method call, renamed something, or changed code that no user could see.
Most of these tests had one thing in common: every dependency was mocked. The tests knew exactly how the code worked inside, but they knew very little about the final result.
Mocks often test how the code works. Fakes let you test what the code does.
