Some people are strongly against this practice of "curl https://installerscript | bash", but how is this different from downloading and running a setup.exe from a random website?
PSA: The fact that the IPCC no longer considers the worst case scenario of global warming, doesn't mean that the scenario was wrong. We're no longer heading there due to the action we took against climate change. If we took even more action, we could probably get rid of the second worst case etc.
A new EU-based social network called "W Social" has been announced. Some people make fun of initiative and say it's doomed to fail, especially since it requires user verification with a real ID. But interestingly, it's built on AT Protocol, meaning it will be compatible with Bluesky. Which also means, through bridging, it will connect to Mastodon and the Fediverse.
Feature request for @hetzner@mastodon.hetzner.social: S3 access keys with permission limited to a subset of my buckets. Any chance we'll get this?
Unfortunately, too many people treat coding-LLMs like a "temporarily embarrassed superintelligence". They will ignore any shortcomings under the assumption that one of the next versions will be able to fix them, so they don't matter. They will treat even mentioning shortcomings as "a coping mechanism by software engineers afraid to lose their job".
If you want to create a new #opensource project and your first step is creating a repository on #GitHub, why not stop and create it on @Codeberg@social.anoxinon.de instead?
RE: https://mitchellaneous.net/2026/04/03/143054.html
Typical nostalgia. I would bet some money that the person who posted this was 16-25 years old in 1993.
With #AI, it became too easy to create and publish low-quality code. It may not matter if you just "vibecode" something for yourself, but if contributions to #opensource are such low-quality "slop", it's a problem for the project owners. Here's an interesting article about the problem: https://redmonk.com/kholterhoff/2026/02/03/ai-slopageddon-and-the-oss-maintainers/
There should be a name for a phenomenon that I think is related to the Dunning-Kruger effect: people see an AI model finding a solution that's smarter than what *they* could have done, so they assume the AI model finds solutions that are smarter than *everyone's*.