exo.cat is a Fediverse instance that uses the ActivityPub protocol. In other words, users at this host can communicate with people that use software like Mastodon, Pleroma, Friendica, etc. all around the world.
This server runs the snac software and there is no automatic sign-up process.
A quite nuanced expert investigation into possible #activist applications of #AI ☺️🤖 https://www.juniperbevensee.com/musings/does-ai-suck #LLMs
Personally I am in the anarchic "hate it" camp but even so, https://bsky.app/profile/juniperbevensee.bsky.social illuminates a lot about how bias and censorship are encoded into #ChatGPT & other commercial tools.
Are you still on #Spotify?
Spotify’s CEO Daniel Ek has raised €600M for his new startup, which is developing AI TECH FOR WAR. Ek still owns 9% of Spotify, but has 37% voting control. His net worth went from $2.5B to $10B in the last two years alone, on the back of paying musicians a pittance in royalties.
And don’t forget:
• Spotify spent $250M of your subscription dollars to invite Joe Rogan to spew his disinformation on their platform.
• They’re still trying to embrace and extinguish Podcasts.
• They’re developing in-house, AI-generated “music” so users will play them (royalty-free) instead of music created by humans (who demand royalty).
And now, he’s using his wealth, created by your subscriptions, to fund tech that will use AI to literally murder humans in war.
Stop funding him. Quit Spotify now.
#QuitSpotify #music #ai #militaryTech #techbro
“Spotify’s CEO invests $1 billion into an AI military start-up — and musicians are fuming”
A well considered deep dive into #AI's (#GANs, specifically) impacts on the #environment here https://mail.cyberneticforests.com/a-personalized-ecology-of-ai/
BUT-
I wish folks would talk about sunk costs (training the models, building data centres in the first place) too. Without allocating a fraction of those costs among an anticipated collective total number of queries / uses, we're coming up short.
Four Construction Vehicles Burned at Data Center in Georgia
"At the end of March, some entered the construction site of the data center hypercampus east of Atlanta, Georgia and burned 4 construction vehicles. This was done to oppose data centers that quickly exhaust the limited water supply of the region. [..] This specific data center has aimed at facilitating AI connection for local companies. #AI not only use more water than other computer processes, but is quickly becoming a way for the state to increase its intrusion in our personal life. And more and more are used by police and state forces to monitor those who oppose them. The people who burned these machines do not believe that our thoughts and movements should be used for profit or incarceration."
Avui un dels nostres serveis essencials, https://farga.eXO.cat, ha tornat a ser víctima de "les Intel·ligències Artificials".
Del que no es parla mai, és que a part dels consums directes en els seus Centres de Processament de Dades (per saber d'això us recomanem @tunubesecamirio@masto.es!), aquests sistemes d'entrenament generen MOLTES despeses energètiques, de hardware, de temps humà, i de recursos computacionals en la infraestructura de qualsevol persona o entitat amb qualsevol cosa a internet.
Als adjunts veureu la ocupació del processador a https://farga.eXO.cat i com baixa de seguida que hi fem una actuació.
La farga és un servei correctament dimensionat per les nostres necessitats i les d'alguns col·lectius que la fan servir (com ara @liberaforms@mastodont.cat).
Però quan vénen "les IAs" a fer peticions indiscriminades, saturen aquests sistemes, amb un ús anòmal.
Feia uns anys d'això en dèiem #DDoS, o atac Distribuït de Denegació de Servei, ara sembla estar normalitzat en el nom de la "#IA".
Quan feu servir aquells serveis, sigueu conscients que, a part de ser l'equivalent a llençar galledes d'aigua pel vàter, també és com forçar a les teves veïnes a tenir tots els llums de casa oberts, sense cap tipus de justificació.
No és coincidència que #DrewDeVault, publiqués això justament fa dos dies: https://drewdevault.com/2025/03/17/2025-03-17-Stop-externalizing-your-costs-on-me.html
Si voleu més detalls de l'incident, són públics, aquí: https://agora.exo.cat/t/2025-03-19-0035-incident-en-farga-per-indexadors-ia/402
I recordeu que per fer aquesta feina possible, l'eXO necessita suport, això també inclou l'econòmic, podeu associar-vos i implicar-vos-hi! https://exo.cat/associar-se/
Sabot in the Age of AI
Here is a curated list of strategies, offensive methods, and tactics for (algorithmic) sabotage, disruption, and deliberate poisoning.
🔻 iocaine
The deadliest AI poison—iocaine generates garbage rather than slowing crawlers.
🔗 https://git.madhouse-project.org/algernon/iocaine
🔻 Nepenthes
A tarpit designed to catch web crawlers, especially those scraping for LLMs. It devours anything that gets too close. @aaron
🔗 https://zadzmo.org/code/nepenthes/
🔻 Quixotic
Feeds fake content to bots and robots.txt-ignoring #LLM scrapers. @marcusb
🔗 https://marcusb.org/hacks/quixotic.html
🔻 Poison the WeLLMs
A reverse-proxy that serves diassociated-press style reimaginings of your upstream pages, poisoning any LLMs that scrape your content. @mike
🔗 https://codeberg.org/MikeCoats/poison-the-wellms
🔻 Django-llm-poison
A django app that poisons content when served to #AI bots. @Fingel
🔗 https://github.com/Fingel/django-llm-poison
🔻 KonterfAI
A model poisoner that generates nonsense content to degenerate LLMs.
🔗 https://codeberg.org/konterfai/konterfai