AI Agents for macro servers.
Posts by Raúl A. Astruga
Fomoria is in its last, trembling days on Kickstarter. Catch it before it sinks entirely into the dark earth.
fomoria.com
Commuting - main theme
Best Star Trek crew, wrong answers only
Qué lujo Peter Svidler
Woke up dead
Monorracha
Capture of a email inbox with a message from chess.com with the subject "You are in a 1 day streak!" in Spanish
Monostreak
Counterproductive if you look for peace and stability, productive if you want to redraw the Islamic Republic's jerarquies, and stablish a full radical goverment (even for the Islamig Republic's standards) invested in prolonging
the military confrontation, so you can justify a prolonged war, lol
KRAA: A short atmospheric game about a murder of crows in an uncertain world.
t.co/nqx0OgRaNT
Please help spread the word #medievalsky!
Funded PhD position on medieval literary and artistic representations of disability as part of a Vidi project I will be leading that is funded by the Dutch Research Council:
www.academictransfer.com/nl/jobs/3592...
i just played Many Nights A Whisper on the strength of @mathewkumar.com's short summary:
"I knew it was simply not going to be matched … thoughtful, paced well, … builds towards an unforgettable moment that you could only do in a video game … just buy it."
deconstructeam.itch.io/many-nights-...
Virtue and a Sledgehammer is a narrative experience.
Who would have thought that so much narrative could be delivered through a hammer?
Wishlist this narrative game:
store.steampowered.com/app/3743260/...
Lo único claro en todo este lío es el nacimiento –en la política, pero también en sus contrarios: la guerra cultural, el nihilismo– de la némesis de Trump: Sánchez
@guillemmartinez.bsky.social ✍️
"Lo de las catástrofes"
ctxt.es/es/20260301/...
Si se entrenara al modelo para decir "honestamente, no estoy seguro" o "no lo se", aproximadamente el 30% de las peticiones y preguntas recibirían esa respuesta, lo que choca con el mercadeo de la Infalible Bola 8
Los LLM entrenados con modelos de IA siempre alucinarán. Matemáticamente demostrado. Los modelos y bancos de pruebas castigan la duda e inducen al modelo a contestar con total confianza incluso cuando su sistema estadístico no le permite predecir la siguiente palabra con certeza.
the art of diplomatic menaces has decayed
JACKIE JAM #3 HAS OFFICIALLY STARTED!
Theme: Signal
You have 10 days to complete the jam! Good luck everybody!
itch.io/jam/jackie-j...
Absolute Martian Manhunter issue 12 cover. Our. Last. Issue. #absolutemartianmanhunter #absolutecomics #dccomics #seriesfinale @denizcamp.bsky.social
El inquietante y opresivo futuro que Marcos Prior dibuja no resulta tan distante como cabría esperar, con personajes y titulares que podrían ser próximos, la manipulación informativa, el monopolio informativo o el poder económico. Muy recomendable @editorialfinestres
There's no dilemma, is a tug-of-war between a model that is proven to give peace and prosperity in Europe or to return to a 1913 kind of status quo.
¿Otra vez? Guardiola tiene perfil de veleta pero de las que siempre apuntan a la ultraderecha.
Godot's GitHub has increasingly many pull requests generated by LLMs and it's a MASSIVE time waster for reviewers – especially if people don't disclose it. Changes often make no sense, descriptions are extremely verbose, users don't understand their own changes… It's a total shitshow. #godotengine
As I keep saying, people who claim there's emergent consciousness in LLMs either don't understand transformer architecture or they don't understand the basics of what consciousness is (or even *cognition* since consciousness is a hard problem), and likely they don't understand either.
It's 968 AC.
An artificial sun shines down upon Doskvol. Autopods cruise down rows of peaceful, prosperous houses. Beneath it all, strange actors and twisted conspiracies work their shadowy agendas.
Blades '68. The future is now. Don't miss it:
www.backerkit.com/c/projects/e...
Moonlighter!