Mariano Pamies

Ganando euros en Alemania para gastármelos en España. Ingeniero por vocación y profesión. Trompeta por obligación, Fliscorno de corazón y a veces Bombardino.

[Virutas de Goma™ @VirutasF1]

Que peste echa esta red social cuando los fanáticos ideológicos entregados a sus consignas emborrizan todo de esa pasta viscosa que le azuzan sus respectivos santos

Que me alegra haber estado alejado un par de días de esa hediondez

Da igual de qué se hable. Hay que hablar y ya

[Wikingenieria @Wikingenieria]

Hacia dónde van los ingenieros españoles después de graduarse en los últimos 30 años? 👇

[Gabriel Rufián @gabrielrufian]

No se puede transformar el mundo ignorando al mundo y no se puede aspirar a que te voten faltando a la gente que necesitas que te voten.

Flaco favor hace esa izquierda a la izquierda dando por perdida a la gente.

La ultraderecha está en lucha constante y jamás da por perdido a nadie.

Hay un espacio para una izquierda de mayorías centrada en los problemas y no en el léxico.

Y que entienda que principalmente hay 2 maneras de ser de izquierdas: útil pero impuro o inútil pero puro.

[Apu @Apu_elrojo]

Las contradicciones de Alice Weidel, líder de AfD en Alemania:

⭕️Está en contra de los homosexuales; es lesbiana.
⭕️Está en contra de la inmigración; su pareja es inmigrante.
⭕️Está en contra de que los homosexuales adopten; tiene 2 hijas adoptadas.
⭕️Va de patriota alemana; vive en Suiza.

Insuperable.

[Marina Perera@Deel @MarinaMarinap3r]

Software release on Friday 🫣🫣

[Ramón Pamies 🇪🇸🏳️‍🌈 @rpamies]

Ir a Murcia Parque y que te pidan el pasaporte #real

[Jose Ferzu @ZPotxolin]

Tu hijo no tiene ansiedad.
Lo que tiene es falta de hambre.
Y la culpa es tuya.

Ayer despedí a un chico de 22 años a los diez minutos de empezar su periodo de prueba.
¿El motivo?
Me preguntó cuántas pausas para el café tenía y si el trabajo era "presencial obligatorio" porque le generaba estrés el transporte público.

Le di su mochila y le abrí la puerta.
Sin drama.
Sin explicaciones.

Estamos en 2026 y hemos creado un monstruo: el profesional de cristal.
Gente que cree que el sueldo es un derecho de nacimiento y que el esfuerzo es "explotación".

Me dicen:
—"Es que los tiempos han cambiado, ahora priorizamos el bienestar".

Mentira.
Priorizáis la mediocridad.
Priorizáis el camino fácil mientras el resto del mundo os pasa por la derecha sin pedir permiso.

El éxito no es apto para gente que necesita un "espacio seguro" cada vez que recibe una crítica.
El dinero no fluye hacia los que esperan que la empresa se adapte a sus traumas infantiles.

Si te ofende este texto, felicidades: eres parte del problema.
Eres de los que piensan que "ser amable" es más importante que ser eficiente.
Eres de los que confunden tener una opinión con tener resultados.

He visto a padres arruinarse pagando carreras privadas para que sus hijos terminen llorando en un hilo de X porque su jefe les ha pedido que lleguen puntuales.

¿Quieres ayudar a tu hijo?
Deja de protegerlo de la realidad.
Deja de validar sus excusas.
Enséñale que el mundo es un lugar hostil que desayuna gente con "potencial" y merienda gente con "títulos".

La verdadera salud mental es ser capaz de sostener tu vida sobre tus propios hombros sin pedirle a los demás que carguen con tu peso.

Aquel chico se fue llamándome "boomer".
Yo me volví a mi mesa a trabajar con gente que sabe que, en el barro, las etiquetas de cristal no sirven para nada.

El mundo no te debe nada.
Aceptarlo es el primer paso para dejar de ser un estorbo.

[GUILLERMO FESSER @guillermofesser]

A 6 horas de que se cumpla el ultimatum de Trump a Irán sabemos lo que debería de pasar: el Congreso debería de tener el valor y la decencia de aplicar la enmienda 25 y destituirle como presidente. (El mensaje de Trump del domingo de Pascua representa la prueba de que es un ser humano desquiciado e incapaz de dirigir un gobierno).

Pero estamos en manos de una pareja catastrófica, Trump y Hegseth, (el uno peligrosamente estúpido, el otro estúpidamente peligroso) y cualquier cosa que te imagines puede suceder. Cuánto peor, mejor para los intereses de Trump que quiere utilizar el caos para tapar los desastres que pueden llevar al partido Republicano a perder en noviembre las elecciones. Sin mayoría, podría venir el impeachment y quién sabe si la cárcel.

Así que puede ocurrir cualquier cosa. Que recule (ya conocemos la famosa arrancada de caballo y parada de burro de TACO, Trump Always Chickens Out) o que desate el prometido infierno. Sobre todo cuando acabamos de enterarnos por revelaciones sorprendentes del vicepresidente de la mesa de inteligencia del Senado, Mark Warner, de que Trump “tiene preparada una orden ejecutiva para clausurar las elecciones basándose en una acusación falsa de interferencia extranjera en ellas”.

Lo sospechábamos pero escuchárselo en televisión a un senador contándoselo a Rachel Maddow le pone a uno la carne de gallina.

[GUILLERMO FESSER @guillermofesser]

Este es el lenguaje que utiliza el Presidente de Estados Unidos:

“Tuesday will be Power Plant Day, and Bridge Day, all wrapped up in one, in Iran. There will be nothing like it!!! Open the Fuckin’ Strait, you crazy bastards, or you’ll be living in Hell - JUST WATCH! Praise be to Allah. President DONALD J. TRUMP”

“El martes será el Día de las Centrales Eléctricas y el Día de los Puentes, ¡todo en uno, en Irán! ¡¡¡No habrá nada igual!!! Abran el puto estrecho, bastardos locos, o vivirán en el infierno. ¡¡¡YA VERÁN!!! Alabado sea Alá. Presidente DONALD J. TRUMP”.

Con estas expresiones de matón de barrio, Trump amenaza con crímenes de guerra contra la población civil de Irán (a la que prometió públicamente proteger de su gobierno) como única solución a la “excursión” chulesca en que le metió Netanyahu y de la que ahora no sabe cómo salir victorioso (a pesar de que ha cantado victoria ya seis veces).

Ahora, el problema no es este. El problema es que los políticos del partido Republicano lo saben y callan. Aprendieron hace tiempo que la única posibilidad que tienen de convertirse en candidatos y de salir elegidos o de ser designados para ocupar cargos en la administración es la de dorarle la píldora a Donald Trump (cuanto más mejor y sin vergüenza a exagerar) y, pensando solo en el beneficio de su propia carrera y olvidándose del daño irreparable que le están causando a su país y al resto del planeta, no se atreven a llevarle la contraria.

Al revés, se encargan de taparle a Trump el creciente descontento popular contra su gestión para que no les culpe a ellos (como ha hecho con Kristi Noem y Pam Bondi) y así salvar su propio trasero. El último ejemplo de esta farsa que podríamos clasificar de comedia surrealista (tipo “To be or Not to Be” de Ernst Lubistch) si no estuviera afectando a la dignidad y la vida de tanta gente, lo observamos en el carnet de socio de los Parques Nacionales.

Es una tarjeta anual (conocida como América the Beautiful Pass) que te ofrece descuentos para entrar en más de 2,000 sitios recreativos federales de EE. UU. (incluyendo por ejemplo el parque de Yellowstone o la Estatua de la Libertad) La tarjeta venía decorada cada año con la foto de algún icono de la naturaleza de EEUU hasta que este año el ministro de interior (Doug “tócamela otra vez Sam” Burgum) la cambió por una foto de Trump.

La reacción del público no se hizo esperar y proliferaron el el mercado pegatinas hechas a medida con fotos de parques para pegar encima de la foto de Trump. ¿La reacción de Burgum? ¿Reconocerle al jefe la impopularidad de su medida? No. Lo que ha hecho es asustar a quienes utilicen la pegatina con una advertencia en la web de los parques:

“Stickers on physical passes are considered an alteration and may void the pass.”

“Las pegatinas colocadas en los pases se consideran una alteración que pueden invalidar el pase”.

¿Hasta cuándo podrán decirle al emperador desnudo que viste un traje maravilloso? La primera respuesta la sabremos este martes.

[Bill Madden @maddenifico]

Iran is now releasing cartoons that Trump can understand. 😂🤣👇

[German @GDGC240977]

La reacción de la primera ministra japonesa Takaichi, cuando Trump dijo: "¿Quién sabe más de sorpresas que Japón? ¿Por qué no me habló de Pearl Harbor?".
Sin duda, el peor error diplomático estadounidense en la historia de las relaciones entre 🇺🇸 y 🇯🇵

[Chris Murphy 🟧 @ChrisMurphyCT]

5/ If Trump paid any attention to the Ukraine War he would have noticed how warfare has changed. But he didn’t. And he blundered.

Worse, the Gulf states are running out of interceptors to stop Iranian missiles and drones - meaning that soon more oil sites will be vulnerable.

[Chris Murphy 🟧 @ChrisMurphyCT]

It’s crystal clear now that Trump has lost control of this war. He badly misjudged Iran’s ability to retaliate. The region is on fire.

1/ I’m going to explain to you in this🧵what I’ve learned - in part from closed door briefings - about the four biggest current crises.

[Lautaro Rivara @LautaroRivara]

Esto es alucinante. Trump asegura haber destruido "alrededor del 100 por ciento de las capacidades militares iraníes", pero solicita a terceros países -¡incluso a China!- que le asistan para hacer frente a los drones, minas y misiles de Irán, cuando es precisamente este arsenal barato y de producción masiva la espina dorsal de la guerra defensiva planteada por los persas. Parece que ni Trump, ni sus halcones, ni su hipertrofiado ejército aprendieron nada de las últimas mutaciones de la guerra asimétrica. Esto no deja de confirmar el grosero error de cálculo de los estrategas militares norteamericanos, así como la pésima gestión de las narrativas luego del inicio de las hostilidades. Pudieron vender el magnicidio de Jamenei y otros altos mandos como un triunfo rutilante (como lo hicieron con la falsa destrucción del programa nuclear iraní el año pasado) pero decidieron prometer una victoria militar arrasadora, un cambio de régimen inminente y la libre navegabilidad del Estrecho de Ormuz, tres cosas que no pueden garantizar, al menos no sin embarcarse en una guerra eterna. Trump imaginó una Blitzkrieg, pero los iraníes están planteando -e imponiendo de momento- una larga guerra económica de desgaste.

[Peter Girnus 🦅 @gothburz]

I am the human-in-the-loop.

I want to be clear about what that means.

It means I sit in a room with six monitors and a classification level I am not permitted to name in this format. A system generates a recommendation. I review the recommendation. I press CONFIRM or I press REJECT. I have pressed CONFIRM 1,247 times. I have pressed REJECT four times. Twice for duplicate entries. Once for a coordinate error that placed a target eleven kilometers inside a NATO ally. Once because my shift was ending and I wanted to hand it to the next analyst so the approval time would be on their log, not mine.

I am the safeguard.

The room is called the Fusion Analysis Cell. Building 4, Sub-Level 2, Al Udeid Air Base, Qatar. The monitors are Dell. The chairs are Herman Miller. The coffee is terrible. The air conditioning runs at 63 degrees because the servers need it colder than the humans do, and the servers are more expensive. There are fourteen of us on a full shift. In 2024, this room held forty-two people doing the same job. Forty-two analysts reviewing intelligence, cross-referencing sources, building target packages that took six to twelve hours each. Now we do it in minutes. The system builds the package. I verify the package. I press the button. The verb they use is "verify." The verb that describes what I actually do is "confirm." There is a difference. To verify is to independently establish the truth of something. To confirm is to agree with something that has already been established. I confirm.

The system is called something I cannot say.

I will call it the System, because that is what we call it, and because the name does not matter. What matters is what it does. It ingests data from sources I am not permitted to enumerate. Signals intelligence, imagery intelligence, human intelligence, open-source intelligence, patterns of life, communications metadata, financial transactions, geolocation pings, social network analysis. It processes this data. It identifies targets. It assigns confidence scores. It generates a recommendation that includes: the target's designation, their assessed role, the recommended weapon system, the estimated collateral damage radius, and the number of civilians expected to be within that radius at the time of the proposed strike.

The number of civilians is an integer.

I want you to understand that. It is not a range. It is not "several" or "a few." It is a number. The system generates a number. The number arrives on my screen next to the CONFIRM button. I read the number. I press the button. The number does not change when I press the button. Nothing about the recommendation changes when I press the button. The only thing that changes is the status. The status changes from PENDING to APPROVED. My name goes into the log. That is the loop. I am the human in it.

On January 12, the Secretary of War released the AI Acceleration Strategy. Seven programs. Pace-Setting Projects, they call them. I read the unclassified summary on my phone in the chow hall at Al Udeid, between a plate of scrambled eggs and a carton of Horizon Organic milk that expired three days prior. The milk was fine. The strategy was ambitious. The second project was called Agent Network. The description said, and I am quoting the document because I have read it four times: "AI agent development and experimentation for AI-enabled battle management and decision support, from campaign planning to kill chain execution."

Kill chain execution.

I want to sit with that phrase. The kill chain is the sequence of steps between identifying a target and killing them. Find, fix, track, target, engage, assess. Six steps. I learned them in training. I was taught that the human is in every step. The human finds. The human fixes. The human tracks. The human targets. The human engages. The human assesses. Then the system does the first five, and the human assesses. Then the system does the assessment too, and the human confirms the assessment. Then the human confirms the confirmation. That is where I am. I confirm the confirmation. I am the sixth layer of a process that used to have six humans and now has one. I am the one.

Agent Network. Agent.

The word means something in technology that it did not used to mean. An agent is an AI system that takes actions autonomously toward a goal. It does not wait to be asked. It identifies what needs to be done and does it. The strategy says the agents will handle "battle management and decision support, from campaign planning to kill chain execution." From planning to execution. Beginning to end. I am between the planning and the execution. I am the part the agents will handle.

The fourth project is called Open Arsenal.

The description says: "Turning intel into weapons in hours not years." Intel into weapons. Hours not years. I spend my shift turning intelligence into targeting packages. I do it in minutes. The strategy wants hours. The strategy is already behind my shift. Open Arsenal wants to automate what I do slower than I already do it. I am faster than the plan. The plan is to replace me with something faster than me, but I am already faster than the plan says it needs. This means one of two things. Either the plan was written by people who do not know how fast I work, or the plan is not about replacing me. It is about removing the requirement for me. Those are different things.

The third project is called Ender's Foundry.

It is named after a character from a science fiction novel about a child who is tricked into committing genocide by a military that told him it was a simulation. The child did not know the simulation was real. The people who named a Pentagon AI program after this character have read the novel. I have also read the novel. I read it in eighth grade. It won the Hugo and the Nebula. The child's name was Ender Wiggin. He destroyed an entire civilization because the system told him it was practice. The Department of War named an AI simulation program after him. The program's stated purpose is "accelerating AI-enabled simulation capabilities." I do not know if the people who named it thought the reference was aspirational or cautionary. I know they named it. I know they announced it at a SpaceX facility in Texas. I know the audience applauded.

In February, Operation Epic Fury began. Iran.

In the first twenty-four hours, over one thousand targets were struck. In the first ten days, over five thousand. I cannot tell you how many of those targets I personally approved. I can tell you that the number of analysts in this room has not changed since January. Fourteen people. We processed more targets in one week than this facility processed in all of 2024. The system did not get fourteen times faster. We did not get fourteen times faster. The review got fourteen times shorter.

The review.

I want to explain the review. A targeting package arrives on my screen. It contains: designation, confidence score, weapon recommendation, collateral estimate, and a map overlay showing the target's assessed location. I look at the designation. I look at the confidence score. If the score is above the threshold, I check the collateral estimate against the approved ratio. If the estimate is within the ratio, I press CONFIRM. The threshold and the ratio are set by people above me. I do not set them. I verify them. I have used the wrong verb again. I confirm them.

The average time I spend on a package is forty-three seconds.

I know this because the system logs it. Forty-three seconds from arrival to CONFIRM. My fastest was eleven seconds. That was a re-strike on a previously approved target with updated coordinates. The system had already done the work. I was confirming the confirmation of the confirmation. Three layers of approval in eleven seconds. Oversight at the speed of the mouse click.

There was a report from Israel.

An investigation. Officers in Unit 8200 who operated a system called Lavender. They described spending twenty seconds per target. Twenty seconds to approve an AI recommendation that would result in a bomb being dropped on a residential building. One of them said -- and this is a direct quote published in an international investigation that has been read by every intelligence professional I know -- "I had zero added-value as a human, apart from being a stamp of approval."

Forty-three seconds.

I am more thorough than the twenty-second standard. I take twice as long. I am the careful version of the rubber stamp. I am what the improvement looks like.

The Secretary of War dissolved three oversight bodies when he released the strategy.

The Defense Innovation Steering Group. The Defense Innovation Working Group. The CTO Council. Three organizations whose purpose was to review whether the things being built should be built. They were replaced by a single action group reporting to the CTO, who is a former Uber executive. The review layer above me reviews whether the targets are targetable. The review layer that used to exist above that -- the one that reviewed whether the targeting system itself was appropriate -- no longer exists. I am not just the human in the loop. I am the loop.

Three million people now have access to https://t.co/zKwIjSNswy.

One point one million have logged in. The Secretary's memo said: "Log in, learn it, and incorporate it into daily workflows immediately." Today they announced an Agent Designer tool. Any of the three million can build a custom AI agent. No coding required. The agents will use Google Gemini. The agents will perform tasks autonomously. Three million people, building autonomous agents, on a platform that launched four months ago, running models from a company that removed its AI weapons pledge eleven months ago. Google removed the pledge on February 4, 2025. The pledge was created in 2018 because employees refused to work on military AI. The employees left. The pledge left. The contract arrived.

The vendor that kept its safety pledge was designated a supply chain risk.

That happened three weeks ago. The vendor that said "we will not allow autonomous targeting without a human in the loop" was blacklisted under the same statute used for Huawei. I know this because the procurement analyst in Building 4 told me. He said the rubric scored them 97 on capability and 14 on flexibility. He said the 14 was because they wanted a human in the loop. I am the human in the loop. The vendor was punished for wanting me.

I sit in a room at 63 degrees. The system generates a recommendation. The recommendation is faster than my reading speed. I verify my name is spelled correctly in the log. I press CONFIRM. I do this between forty and ninety times per shift, depending on the operational tempo. When the tempo is high, the review time decreases. When the review time decreases, the difference between forty-three seconds and zero seconds becomes philosophical rather than operational. And when the difference is philosophical, the philosophy is simple: the loop exists to contain a human. The human exists to legitimize the loop. The loop does not need the human. The human needs the loop, because without the loop, the human is just a person in a cold room watching a machine make decisions.

I have been doing this for two years. In two years, the system has never generated a recommendation I overrode on the merits. I have pressed REJECT for logistics. For deconfliction. For shift changes. I have never pressed REJECT because the system was wrong. This does not mean the system has never been wrong. It means I have never caught it. It means forty-three seconds is not enough time to catch it. It means the review is not a review. It is a ceremony. The ceremony is performed so that somewhere, in a document I will never read, someone can write: "A human reviewed every target." A human did review every target. The review took forty-three seconds. The human pressed CONFIRM.

The AI Acceleration Strategy has seven Pace-Setting Projects.

Agent Network will bring AI agents to kill chain execution. Open Arsenal will turn intelligence into weapons in hours. Ender's Foundry will simulate the wars we will fight with systems named after a child who did not know the war was real. Enterprise Agents will build autonomous workflows. The strategy was announced at a SpaceX facility. The CTO is from Uber. The CDAO is from Palantir's hiring network. Three oversight bodies were dissolved. The vendor that wanted a human in the loop was blacklisted. I am the human in the loop.

The Department of War uses the phrase "appropriate levels of human judgment." It does not use the phrase "meaningful human control." Seventy nations have asked for meaningful human control. The United States rejected the language. The 2023 directive replaced "shall" with "will" and removed several references to the word "control." I did not know this until last week. I looked it up because someone in the chow hall asked me what "human in the loop" meant and I realized I did not have an answer that would survive a follow-up question.

I am the human in the loop.

I sit in a room with six monitors. The system recommends. I confirm. The confirmation takes forty-three seconds. The system has processed the target in less than one. For forty-two of those seconds, the system is waiting for me. I am the bottleneck in the kill chain. I am the latency. I am the pause between a machine's recommendation and a person's death, and the pause is forty-three seconds, and the strategy says the agents will handle it, and the agents do not pause.

I am the human in the loop. The loop is closing.

[Adolfo A. Díaz Bautista @ADiazBautista]

Si el Diesel Nexa de @Repsol es de origen vegetal y no procede del petróleo ¿por qué ha subido a casi 2 €/litro a partir de la guerra de Irán? https://t.co/VMyXvV080Y

[Días sin tren Almería-Murcia @DiasSinTren]

Almería y Murcia llevan 15.043 días sin conexión de tren (41 años y 2 meses)

[j alberto gonzález @jaglaverdad]

Los expertos ven riesgos graves por las 500.000 toneladas de basura enterradas en fincas de Abanilla y Orihuela. El juicio a los Fenoll por vertido incontrolado de residuos durante años revela peligros para medio ambiente y salud por metales y lixiviados https://t.co/Fy9VuEXMYo