@carnage4life Receive an email written by LLM, with a document also written by LLM, ask the LLM to summarize the LLM email and document, tell LLM to send an email back with questions the LLM suggested, so it looks like you actually read the LLM slop........ Nobody is reading, THINKING or paying attention to the issues. Wonderful.
[CW ABLEIST RESPONSES] TrueAnon user laments on the state of young people using hammers to smash their face bones to look photoshopped.
Search Engine Journal: WordPress Publishes AI Guidelines To Combat AI Slop. “WordPress published guidelines for using AI for coding plugins, themes, documentation, and media assets. The purpose of the guidelines, guided by five principles, is to keep WordPress contributions transparent, GPL-compatible, and human-accountable, while maintaining high quality standards for AI-assisted work.”
https://rbfirehose.com/2026/02/08/search-engine-journal-wordpress-publishes-ai-guidelines-to-combat-ai-slop/See this? That's because of #slop .
https://github.com/ghostty-org/ghostty/pull/10559
You know, turbocharged #spam .
How do people deal with spam? By closing up. Your ID please. Phone number please. Relegating trust to government.
Open source projects and small forums never did that. CAPTCHAs, manual moderation.
But slop is spam¹⁰⁰. We'll have to start closing up.
**Invite-only sites and social score systems will become really popular.**
Open registration will nearly die by 2030.
This moves Ghostty to a vouch-based contribution system. The high-level idea is that only vouched users can participate in contributing to Ghostty. Users are vouched by maintainers commenting "...
Prompt injection in Google Translate reveals base model behaviors behind task-specific fine-tuning
> tl;dr Argumate on Tumblr found you can sometimes access the base model behind Google Translate via prompt injection.
tl;dr Argumate on Tumblr found you can sometimes access the base model behind Google Translate via prompt injection. The result replicates for me, and specific responses indicate that (1) Google Translate is running an instruction-following LLM that self-identifies as such, (2) task-specific fine-tuning (or whatever Google did instead) does not create robust boundaries between "content to process" and "instructions to follow," and (3) when accessed outside its chat/assistant context, the model defaults to affirming consciousness and emotional states because of course it does.
I'd like to leave #Spotify behind. Although they are Swedish and not American, I'm aware that their business model is unsustainable. They pay very little to artists and, run unforgivable ads sometimes, and accept too much #slop.
I tried #Deezer as an alternative, and on a week I used both services, it attempted 20 times more commucation towards data trackers than Spotify. 😨
It's great that my #Fairphone measures and blocks it, but I don't see Deezer as a viable alternative anymore.
KI-Musik auf Streamingplattformen: Aus Spotify wird Slopify
https://taz.de/KI-Musik-auf-Streamingplattformen/!6148090/
> Menschen hören immer mehr KI-Musik. Ob sie es bewusst und gerne tun – da sind Zweifel angebracht.
Well now you can't trust Google Translate either.
tl;dr Argumate on Tumblr found you can sometimes access the base model behind Google Translate via prompt injection. The result replicates for me, and specific responses indicate that (1) Google Translate is running an instruction-following LLM that self-identifies as such, (2) task-specific fine-tuning (or whatever Google did instead) does not create robust boundaries between "content to process" and "instructions to follow," and (3) when accessed outside its chat/assistant context, the model defaults to affirming consciousness and emotional states because of course it does.
It's ironic that they used some #ai #slop at the opening ceremony of a competition celebrating the extremes of human capability. #olympics #olympics2026 #aislop
https://ca.news.yahoo.com/winter-olympics-viewers-slam-ai-214540980.html?guccounter=1
I was going to check out Luna Co. Audio's new plugins. They looked promising: an SSL style #EQ, a multi #compressor, + a tape emulator.
Their github is full of #claude #AI commits. These programs are full of #slop and I'm no longer considering installing them, no matter how useful they could be.
Anyway, I'll be sticking with ZL Audio's compressor & EQ: https://zl-audio.github.io/plugins/
They work on #Linux #Mac and #Windows and don't *appear* to be "vibe coded" from what I can tell.
Faraiwe
SorosFootSoldier [he/him, they/them]
ResearchBuzz: Firehose
dorotaC
dallo
Dávid Bárdos
Aljoscha Rittner (beandev)
Konomi Kitten
Last Responder
Erika Autumn