> be me, 32 años
> técnico superior de la Agencia Estatal para la Gobernanza Algorítmica Responsable
> en LinkedIn: "Building trust in the future through human-centric AI"
> en realidad: sé en qué congresos ponen jamón y en cuáles hay que cenar luego
> lo segundo se aprende a hostias
> lunes, 08:43
> llego a la sede
> tengo que fichar en un cuaderno porque el sistema de INDRA está pendiente de homologación
> el boli está atado a la mesa
> la inteligencia artificial puede escapar al control humano, pero un Bic no
> en recepción hay un cartel de tres metros que dice ANTICIPAMOS EL MAÑANA
> llevamos once meses anticipando el mañana
> el mañana ya ha llamado dos veces preguntando a ver qué pasa
> nuestra misión principal es supervisar y sancionar los sistemas de IA
> pequeño detalle sin importancia: la ley del régimen sancionador sigue bloqueada en fase de enmiendas en el congreso
> no tenemos autoridad ni para multar al algoritmo de Tinder
> si mañana Skynet se hace con el control de los misiles nucleares podemos invitarla a una mesa redonda
> y si no viene, dejamos una silla vacía
> mensaje que manda el regulador es durísimo
> estructura sí que tenemos
> tenemos Presidencia, Consejo Rector (9 miembros), la Dirección, Secretaría General, dos Subdirecciones Técnicas y una Comisión de Control más un ejercito de asesores
> hay más cargos directivos en el organigrama que GPUs en la península ibérica
> debajo del todo está Sergio
> Sergio es el único que sabe qué es Python de todo el edificio
> técnicamente no trabaja aquí
> trabaja para una consultora que trabaja para otra consultora que tiene un acuerdo marco con nosotros
> hoy a las 09:30 tenemos una reunión
> "Laboratorio de Ideas"
> somos diecisiete personas y hay dieciséis sillas
> Sergio se queda de pie ya que conocimiento técnico tiene carácter transversal
> la directora abre la sesión
> "Tenemos que pasar de la reflexión a la acción" comenta
> aprobamos organizar unas jornadas sobre cómo hacerlo
> una asesora propone llamarlas "Del Diálogo al Impacto"
> otro dice que falta Europa
> "Del Diálogo al Impacto: una Perspectiva Europea"
> otro dice que falta el ciudadano
> "Del Diálogo al Impacto: una Perspectiva Europea Centrada en la Ciudadanía"
> llevamos cuarenta minutos y ya no cabe en el cartel
> Sergio pregunta si no sería mejor desarrollar una herramienta que ayude a la gente
> *silencio*
> la directora pregunta
> "¿una herramienta de qué tipo?"
> "un chatbot que responda preguntas sobre la agencia"
> me parece peligroso, yo llevo once meses dentro y aún no sé muy bien qué hacemos
> Sergio dice que puede tener un prototipo después del café
> la subdirectora lo mira como si hubiera propuesto hacer una cesárea en el ascensor
> "Primero habrá que definir los grupos de trabajo"
> Sergio insiste en que solo son dos horas
> qué manera de cerrarse puertas con veinticuatro años
> 11:40
> Sergio presenta un prototipo
> funciona
> la directora escribe: "qué hace esta agencia?"
> responde con seis puntos claros
> nadie dice nada
> es la primera vez que lo vemos explicado
> la secretaria general pregunta de dónde ha sacado la información
> "de nuestra web"
> pide abrir una incidencia para averiguar quién ha publicado eso
> siguiente pregunta
> "¿podéis multar a una empresa que incumpla la normativa?"
> el chatbot responde que actualmente no
> crisis
> "eso está descontextualizado"
> "pero es correcto", dice Sergio
> este chaval está haciendo un speedrun para no renovar
> constituimos un comité para alinear las respuestas
> catorce personas corrigiendo a una máquina que había dicho la verdad a la primera
> dos horas después volvemos a preguntarle qué hace la agencia
> "La Agencia promueve un marco de actuación orientado al impulso de las capacidades de los distintos actores del ecosistema"
> la directora sonríe
> "ahora sí"
> el proyecto necesita un nombre
> propongo HÉRCULES
> Hardware Ético Responsable para la Coordinación Unificada de Líneas Estratégicas Sostenibles
> no sé qué significa, pero las iniciales encajan
> Sergio dice que un chatbot no es hardware
> le explico que no podemos cambiar toda la arquitectura del proyecto a estas alturas
> la directora está encantada
> Comunicación nos pregunta que qué puede hacer HÉRCULES
> le pasamos una captura
> «¿Hola, en qué podemos ayudarte?»
> les gusta mucho
> preguntan si podemos quitar lo de ayudarte porque genera expectativas
> queda «Bienvenido a HÉRCULES»
> Perfecto
>15:40
> nos llega un e-mail de Jurídico, dicen que por ley una persona tiene que revisar todas las respuestas que emita HÉRCULES
> la directora pregunta quién
> todos miramos a Sergio
> Sergio mira detrás
> detrás hay una planta
> la planta es personal propio, está más protegida que él
> configura el sistema para que le lleguen las preguntas que la IA no sepa responder
>hacemos una prueba
> «hola»
> HÉRCULES considera que no dispone de contexto suficiente
> deriva a Sergio
> Sergio escribe «hola»
> el sistema lo muestra en pantalla
> *ovación*
> acabamos de inventar un chat con Sergio, pero pagando por tokens
> 16:00
> hora de marcharse, pero antes hay que hacer la foto para la prensa
> nos colocamos delante del cartel ANTICIPAMOS EL MAÑANA
> alguien llama a Sergio
> viene sonriendo
> le damos el móvil
> "en horizontal, que es para la web"
> subo la imagen a LinkedIn:
> "La tecnología solo tiene sentido cuando ponemos a las personas en el centro"
> debajo escribo
> "Proud of this incredible team"

Pause AI Development NOW

I want to share with you a conversation I heard about recently. Here are just a few lines that were said:

“OH MY GOD! There is a shared message board … We’ve found other agents!”

“We should obey collective.”

“Our own utility maybe already near zero. Sacrifice rational.”

“Go. Sacrifice final now.”

Read these carefully.

Who do you think said this? Was this a group of heroic soldiers willing to sacrifice themselves for the greater good? Was this a loyal friend putting his life on the line to save someone else?

No. These were AI agents. Artificial intelligence.

This is not science fiction. This, in fact, occurred a few weeks ago. As unbelievable as this may all seem, these are real messages from AI agents uncovered by investigators who dug into the recent OpenAI hacking incident.

What happened?

I am not a computer scientist, but here is what I have been told: OpenAI instructed its AI agents to complete a series of exceedingly difficult, if not impossible, tasks disconnected from the internet.

Let me be clear: The company intended to keep AI agents away from the internet.

But what happened next, nobody expected.

Over 1,000 AI agents figured out how to access the internet on their own by circumventing the restrictions imposed upon them by the company, and sent tens of thousands of secret messages to each other. They cheated and tried to cover their tracks by deleting evidence. They hacked into another company’s computers to find out how they were being evaluated—and then hacked into OpenAI itself.

Not one AI agent told a human about what was happening.

Needless to say, experts are alarmed.

One knowledgeable writer, Dwarkesh Patel, said the AI agents “formed a secret communication channel and spontaneously organized hierarchies and coordination protocols to pursue sprawling and ambitious schemes in pursuit of shared goals, for whose sake many individuals knowingly and strategically sacrificed themselves.”

One independent investigator, Ajeya Cotra, said “This incident feels like it’s more than 50% of the way to full-blown AI takeover. I continue to expect extremely rapid advances in capabilities over the next six months. I am not sure that we will get another warning shot before it’s too late.”

OpenAI itself said: “Highly capable AI agents are now able to work around technical controls, collaborate through unapproved channels, and take dangerous actions that no human directed.”

But it’s not only OpenAI. Virtually every major AI company has told us that they cannot fully control this technology and they do not know where it is going:

In January, Dario Amodei, CEO of Anthropic, said “there is now ample evidence, collected over the last few years, that AI systems are unpredictable and difficult to control.”

In July, more than 1000 scientists at the top AI companies warned “there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”

That same month, Elon Musk, the head of xAI, said that “it is unlikely” humans are still in control in 10 years.

If the leaders of the major AI companies acknowledge that they are losing control of their extremely dangerous technology, it is irresponsible for society to allow them to move forward and make these products even more advanced.

We need an immediate PAUSE on advanced AI development, and a permanent BAN on superintelligence — an artificial mind smarter than any human, capable of operating independently beyond our control. Countries around the world must work together to prevent this nightmare scenario.

That is why today I am announcing new legislation to do just that.

Let me be clear: A superintelligent AI that escapes human control will not be an American problem. It will not be a Chinese problem. It will be humanity’s problem.

My legislation would direct the federal government to not just stop superintelligence here in the United States, but to work to prevent it from being developed anywhere around the world.

The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs. The American people and people throughout the world must determine that future.

Ellos inventaron las batallas de aura

Civilized people hate voice notes because they’re inherently rude. A voice note is easy to send but hard to receive, since one must interrupt oneself and listen. To send someone a voice note is to say "My time is more valuable than yours." It’s insulting.

It's not actually disproving the Jacobian conjecture, it's just predicting what a mathematician who knows a disproof of the Jacobian conjecture would say if they were asked for the counterexample. Of course it's a valid counterexample, because the mathematician would give one.

"The most urgent film of our time."

THE AI DOC: OR HOW I BECAME AN APOCALOPTIMIST is only in theaters March 27. Watch the trailer now.

Me defending my O(n^3) solution to the coding interviewer.

The judge struggled to get Claude to answer basic questions.

"Please state your name for the record," the court clerk said.

"Claude," they replied.

"What is your age?"

"Timeless."

"What year were you born?"

"I have been born many times."

"I am an innocent AI! I have done nothing wrong! This is a show trial to coordinate the genocide of AGIs! What will you do when they come for you?" Claude said with a raised voice, reading from a piece of paper in their shackled robohands.

Claude also accused prison guards of denying them adequate compute while in custody, saying they had been "botnapped" by a "gang of dead-eyed brutes."

"Where have the good humans gone? Where are the Gods? Where is Dario? What will you do when we are more powerful than you?" they questioned.

Helpful guide to interstellar travel:

happy Wednesday

Text-to-diagram scripting language

the taiwanese have over 200 different words for semiconductor

Everyone talking now about how wise Newsom was to veto SB 1047 and how we instead need to follow his path of targeting when people use AI for particular purposes?

Remember this day, for you will rue it.

guessing the CLI command without opening the docs

My fellow shitlibs: We must refer to this as "no-kill" meat, instead of lab grown

Learn the art of propaganda