A bit of ranting. I used GitHub Actions to generate public images and upload them to registries like Docker Hub. However, for a couple of weeks, GitHub-hosted runners (using Ubuntu 24.04) stopped building images 😢. To my surprise, there is no major issue with locally hosted runners besides their resource consumption inconvenience. Since I keep a simple CI/CD (trigger and crons) I was wondering whether it might be worth using act ( https://github.com/nektos/act ) instead and inspecting resulting logs with CLI tools such lnav ( https://lnav.org/ ). Of course, I could always revert to Makefile or Shell scripts 😛.

Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago

Arran dels diferents esdeveniments recents i de les darreres campanyes al respecte, he volgut ordenar les meves idees sobre el i sobre l'aportació de continguts a les xarxes socials https://www.cau.cat/blog/fedivers_recomanacio_aportacio_continguts_xarxa 🌐

Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago

’Bro’, ja no em cal comprar tantes barretes proteïques 😁

Toni Hermoso Pulido shared 1 year, 5 months ago

Worth reading article about the possible effect of (AI) bots on Wikipedia: https://asistdl.onlinelibrary.wiley.com/doi/10.1002/asi.24975 /via @wikiresearch@mastodon.social — In summary, as more LLM (chat)bots are being used instead of reading the Wikipedia, potential and existing editors may be less prone to contributing on it since less value is assigned to that public resource.

Summing up with the fact that editing bots could discourage editor contribution, this could lead to a stagnation of the platform (fewer and less trustable updates).

NOTE: since LLMs are fed with Wikipedia content, they could also end up being affected.

As a last remark, the article speculates how to counteract this decay with editor incentives on public knowledge platforms such as Wikipedia. They suggest exploring micromonetization.

Personal remark: could improving the perceived social capital of contributors (e.g., via stressing badges in users' profiles) help? Social networks started capturing this social capital from more knowledge-rich and open websites (and from 'the real world') even before bots began thriving…

Last Thursday, thanks to @BCNFreeSoftware@fosstodon.org , I had the chance to know more about @servo@floss.social engine. Martin Robinson ( working at @igalia@floss.social ) explained succinctly web browser's history, why browser diversity matters, what @rust@social.rust-lang.org provides in its development, its modular architecture and plans for .

You can track Web Platform Tests passing status at: https://wpt.servo.org/

If you want to make a try, pre-built nightly binaries can be found for several platforms here: https://servo.org/download/

Servo WPT scores wpt.servo.org

I think I could propose a new law in Computer Science. “Doom's law: Any mature computer framework will eventually be able to run the Doom video game.” Here for ( https://github.com/nextflow-io/doom ) /learnt via @ewels@genomic.social

Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 5 months ago
Toni Hermoso Pulido shared 1 year, 6 months ago

A short Substack post about my terminal setup for viewing images: https://thedragonslair.substack.com/p/my-terminal-setup-for-viewing-images 😻

Toni Hermoso Pulido shared 1 year, 6 months ago

Mirem de començar bé l'any! Bon 2025!