More teams are rethinking cloud-hosted LLMs. agent.cpp provides minimal, high-performance building blocks for agents in C++, built directly around the llama.cpp ecosystem. Designed for running agents locally on your own hardware. Repo: link.mozilla.ai/agent-cpp
David de la Iglesia Castro
@daavoo.bsky.social
Climbing. Surfing. Board gaming. Working from Motorhome. Open-source AI at https://www.mozilla.ai/
Guardrails are vital for AI safety, but hard to choose. Each has its own logic & definition of unsafe. That's why we built any-guardrail, a unified interface to test and switch guardrails seamlessly, making AI safer, more reliable, and easier to build. Read our post: blog.mozilla.ai/introducing-...
Introducing any-guardrail: A common interface to test AI safety models
Building AI agents is hard, not just due to LLMs, but also because of tool selection, orchestration frameworks, evaluation, safety, etc. At Mozilla.ai, we’re building tools to facilitate agent develop...
blog.mozilla.ai
Working with multiple platforms & LLMs, but some models only appear on certain providers? Try any-llm Model Finder! See connected providers via API keys & search models across all of them. It's built on top of our any-llm library. Check out our repo and GIF: github.com/mozilla-ai/a...
any-llm/demos/finder at main · mozilla-ai/any-llm
Communicate with an LLM provider using a single interface - mozilla-ai/any-llm
github.com
ESPAÑA LLEVA SIN LUZ QUINCE MINUTOS. AHORA IMAGINAD GAZA, QUE LLEVA SIN LUZ AÑO Y MEDIO.
🖥️ O galego ábrese camiño na Intelixencia Artificial www.nosdiario.gal/articulo/lin...
O galego ábrese camiño na Intelixencia Artificial
O Proxecto Nós continúa a sumar fragmentos de voz para adestrar ferramentas de recoñecemento automático da fala en lingua galega. O idioma propio da Galiza avanza así cara á súa presenza nos asistente...
nosdiario.gal
Las prisiones españolas usan un #algoritmo sin actualizar desde 1993 que es decisivo para determinar permisos de salida. Además de que, 200 resoluciones en 2024 citan el resultado de este algoritmo para denegar u otorgar permisos, tiene algún que otro sesgo. Aquí, nuestra última investigación👇
Las prisiones españolas usan un algoritmo sin actualizar desde 1993 para decidir sobre permisos de salida
Más de 200 resoluciones de distintas audiencias provinciales dictadas en 2024 citan el resultado de este algoritmo como motivo para denegar u otorgar permisos penitenciarios. La TVR clasifica a los…
civio.es
Meet Any-Agent, a Python library that lets you build and switch across agent frameworks. Try it: github.com/mozilla-ai/a... We built it because we were struggling. Now we’re sharing. 🔹 Unified interface 🔹 Logging with OpenInference 🔹 Built-in eval Blog: blog.mozilla.ai/introducing-...
New report from @greenpeaceeastasia.bsky.social and @digiconomist.bsky.social finds that AI companies like Nvidia are threatening climate progress Climate pollution from AI computer chip production increased by *over 350%* in just one year (That's before they're even used!)
Artificial Intelligence Supply Chain Threatens Climate Progress in East Asia - Greenpeace East Asia
Electricity consumption from the manufacture of artificial intelligence (AI) chips has soared by more than 350 percent worldwide between 2023 and 2024,
greenpeace.org
invariantlabs.ai/blog/whatsap...
WhatsApp MCP Exploited: Exfiltrating your message history via MCP
This blog post demonstrates how an untrusted MCP server can attack and exfiltrate data from an agentic system that is also connected to a trusted WhatsApp MCP instance, side-stepping WhatsApp's encryp...
invariantlabs.ai
Announcing a new event initiative: Feminist AI LAN Party! Katharine and I did a pilot event last year and we're now taking it to @pyconde.bsky.social. We've also open-sourced event kits to make it easy to host your own, including: 💣 hacking LLMs 📑 data development ✂️ zine making feministai.party
You can now register at www.eventbrite.com/e/dair-prese... ⏰Friday April 11 9am-12pm PDT | 12pm-3pm EDT | 6pm-9pm CEST | 7pm-10pm EAT
DAIR Presents: Imagining Possible Futures
At DAIR, we want to encourage imagining otherwise. Join us for two panels and a collaborative zine-making workshop.
eventbrite.com
We're hoping to organize a hopeful virtual event to get us to think our possible technological futures, and have a creative workshop. Stay tuned. In the mean time, read our Possible Futures Series. If you have thoughts on what you'd like to see from the event, comment below.
MANIFESTACIÓN HISTÓRICA POR TIERRA Y MAR Galicia ha hablado: no quiere una nueva fábrica de celulosa ni está dispuesta a que se malgaste el dinero público en proyectos sin licencia social y tan agresivos con su naturaleza, su forma de vida y su cultura. ¡No a la megafábrica de celulosa! ALTRI NON!
Israeli forces have killed Hossam Shabat with a targeted airstrike while he was driving his car in Beit Lahiya. Shabat, who was 23 years old, had become well-known for his reports from northern Gaza amid Israel’s total siege on the region.
Israel Kills Journalist Hossam Shabat, Known for His Reports From North Gaza
“I ask you now: do not stop speaking about Gaza,” he said. “Keep telling our stories — until Palestine is free.”…
truthout.org
Dozens more pregnant and postpartum women have died in Texas hospitals since the state banned abortion, our analysis shows. As the maternal mortality rate dropped nationally, it rose in Texas by 33%.
Texas Banned Abortion. Then Sepsis Rates Soared.
ProPublica’s first-of-its-kind analysis is the most detailed look yet into a rise in life-threatening complications for women experiencing pregnancy loss under Texas’ abortion ban.
propub.li
OpenAI doesn't report a single number for its climate impacts - not anywhere, in any way, shape or form. It sounds kind of obvious but the fact that a company with such incredible electricity hunger isn't disclosing basic information about its energy consumption and associated emissions is v bad..
¿Por qué mantener el contexto histórico y social, la clase, el género, incluso el peinado, y decidir adelgazarlas? Cristina de Tena escribe sobre la historia de Hildegart y la película 'La virgen roja'.
La virgen roja era gorda
Periodismo de calidad, con perspectiva feminista, crítico, transgresor y disfrutón desde 2010
buff.ly
"When trying to develop a measure of intelligence, it’s essential to avoid Goodhart’s law: “When a measure becomes a target, it ceases to be a good measure.” As a community of AI researchers, we really need to figure that one out." @melaniemitchell.bsky.social aiguide.substack.com/p/did-openai...
Did OpenAI Just Solve Abstract Reasoning?
OpenAI’s o3 model aces the "Abstraction and Reasoning Corpus" — but what does it mean?
aiguide.substack.com
✨ Introducing Mozilla.ai Blueprints: customizable AI workflows for developers. 📢 Blog post: blog.mozilla.ai/introducing-... 🔥 First up: Document-to-Podcast—transform documents into multi-speaker audio clips with open-source tools. github.com/mozilla-ai/d... #MozillaAI #AI #OpenSource #Blueprints
Decenas de miles de manifestantes de toda Galicia han colapsado las calles del centro de Santiago, coreando lemas sobre de la defensa del patrimonio natural Los vecinos protestan contra la instalación de la macrocelulosa de Altri, que generará 62 toneladas de residuos químicos tinyurl.com/4wx6r532
And if you want to know more about the economic reality behind running high availability, actually innovative tech… signal.org/blog/signal-...
Privacy is Priceless, but Signal is Expensive
Signal is the world’s most widely used truly private messaging app, and our cryptographic technologies provide extra layers of privacy beyond the Signal app itself. Since launching in 2013, the Signal...
signal.org
“The automated disinterest of algorithmic bureaucracy can transform public services into an aesthetic exercise, prioritizing the symbolic achievement of dashboards and metrics: Your health insurance eligibility decision is no different from AI Slop.” ⬇️⬇️⬇️
Slop Infrastructures 3 & 4
Weaponized Slop 💡This week, I'm trying something new. I'll be e-mailing smaller sections of a single piece, Slop Infrastructures, each day. This includes part three and four. Parts one and two are he...
mail.cyberneticforests.com
📢NEW: 'Open' AI systems aren't open. The vague term, combined w frothy AI hype is (mis)shaping policy & practice, assuming 'open source' AI democratizes access & addresses power concentration. It doesn't. @smw.bsky.social, @davidthewid.bsky.social & I correct the record👇 nature.com/articles/s41...
Why ‘open’ AI systems are actually closed, and why this matters - Nature
A review of the literature on artificial intelligence systems to examine openness reveals that open AI systems are actually closed, as they are highly dependent on the resources of a few large corpora...
nature.com
What would happen if proven open source maintainers built an Agent Framework? We're about to find out...
PydanticAI is here! An Agent Framework designed for production, from the team who created and maintain @pydantic.bsky.social. As some of you will know, I've been working on this for some time, can't wait to see what people build with it. ai.pydantic.dev
New edition of Just enough Internet just landed in inboxes. This one on how non-technical people actually appear to be using generative AI at work. buttondown.com/justenoughin...
FOMO is not a strategy
End of year reflections on how people are using genAI at work
buttondown.com