Gregory's Server
|
Having a choice between different opinions, platforms, and everything is a democracy. Telling everyone about the choice you made is propaganda. That’s simple. Today I was one inch from rating anything online for the first time in my life. Goddamn security through obscurity and user penetration. My mobile app for car charging _asked me to change the password_ right when I was to start changing. Imagine a garage, with very poor internet, I am hanging up with a damn cable in one hand and my belongings in the other—and voilà—change your password now, I am not going to scan the QR code otherwise. As a person who provided hundreds of pull requests to many different repositories in many different languages, I don’t understand what “bus factor” is supposed to mean. There is a code, why would you need anything to understand it? Huh? FWIW, I am subscribed to several people (bots?) who post “art” made by “AI.” I scroll through dozens of generated pics daily. There was not one yet I wanted to print and put on my wall. ——— Tomorrow I’ll drive my friends to David Vela’s exhibition in Zaragoza to acquaint them with the most heartwarming artist of our times. Carrying Gogol’s Petersburg Novels (Петербургские повести, Novelas breves peterburguesas) with many illustrations by Russian artists with me as a present, fwiw. Code Coverage might be indeed a usable metric. Check this “17.1 hits per line” value. The actual efficiency of the code is literally the density of its usage. 49 lines with 17.1 hits per line are a dense implementation of what 838 LoCs could do in a worse code. List of my libraries updated to `Elixir v1.16` - camarero There is a lovely, lived story behind every great painting. There is a clumsy, edited prompt behind every AI odd job. TIL: docker is unable to update itself as an apt package with my nightly cronjob. Funcking unbelievable. I have never seen worse software (even Electron, Flatpak, and Snap are way better.) Most developers struggle to determine whether one needs a test for the newly introduced code. The answer is on the surface: if it makes it easier to write robust code, the test is likely needed. Don’t waste your time otherwise. Your language of choice and the framework you leverage to build your project do not require you to test them. 102% of discussions on social media look like the following arguing about the color of snooker balls. ▸ 68% stand for balls are red; And yes, it does not amount to 100%. Назрел вопрос: вот есть копирайтеры, так? Почему в сети все достойные тексты — написаны людьми иных профессий? Как так получилось, что словарный запас каждого пятого программиста — примерно втрое богаче? Что мешает им заниматься проверкой фактов так же тщательно, как, например, мне? А?  С чего бы человеку, умеющему самостоятельно составлять сложные тексты и обладающему большим словарным запасом, заниматься работой, подразумевающей коверканье чужих текстов? Ящитаю, ситуация более чем закономерна. :)
Effective tomorrow, I modify the license for all my OSS projects to explicitly prohibit the usage for training AI and/or for producing AI-related products. So what? If it’s impossible, don’t. That simple. Do these fucking morons look for empathy now? https://www.theguardian.com/technology/2024/jan/08/ai-tools-chatgpt-copyrighted-material-openai |
Интересно:
1. перевод на какой язык мне предлагает мастодон?
2. справится ли переводческий движок с таким материалом?