ChatGPT vuelve a ser mejor que Claude gracias a la potencia de cómputo de los WC japoneses

Hello to all the monos estocastic and welcome to the podcast favorite of the jurado of Elon Musk against Sam Allman. Because they have to decide something more difficult that can say a jurado, which is, who is more vendahumos, if Elon Musk or Sam Allman. The truth is, I don't like to be in his shoes, but listening to the podcast, a lot of people come to ideas. How are you, Antonio? Ah, Mati, it's a hard work that's the jurado, but I'm going to confess that I have already put the head in the summer. We had the bridges here in Spain, we made a little family escape. And again, I'm going to think about visiting other places, to know other worlds. esa parte de los viajes que tiene de ejercicio de fantasear con que otras vidas es posible, con descabalgarte de la rutina y el día a día y jugar, juguetear a conocer nuevos lugares con otros ritmos, otras comidas y otros lenguajes. Sin embargo, Mati, hay que advertir a la audiencia y además te voy a confesar por primera vez que los viajes también aguardan un lado oscuro, Algo que puede resultar inquietante y que te voy a dar un testimonio familiar real que nos ha sucedido. Verás, Mati, estando de viaje, disfrutábamos del embarcimiento, visitas turísticas, comidas excesivas, el triple desayuno, como son los hoteles. when, in a flora without anticipating it, without anesthesia, my daughter made me one of those revelations that you're going through. There are moments in the life that the existence, everything that you want to know, can be separated in two, because Daniela me looked at fijamente and told me what the following is. Dad, what if we make a TikTok family? In the world, of course, I can't culpar it. In the adolescence, the vision of the world is still affected by the success of the platforms and the llamadas of the viral, the fame and the content meetic. But, of course, in the middle, I broke my world. The electronic plates were moving. And a cold cold air me recorred my back. Even so, Matisse, I can tell you something. I managed to maintain the compost. I was not coming down. I was not being a terror, which was what the situation was. And with much calm, I said, well, let's see what happens to be a family TikTok. and she told me, well, basically we have to do the video of the tele-transport. You are in a place, you are a nice gesture, and of a sudden the entire family appears in another place that is very deseable. Like, of a sudden, you transport a the mountain of Trevi, or Manhattan, or anything like that, that result in a llamative and is a great thing in TikTok. I was there, I was a little bit me bajaron las pulsaciones cogí aire y respiré hondo y me quedé un poco más tranquilo pero ahí todavía me aguardaba la última estocada porque dijo, ah papá hay otra cosa tiene que parecer que somos felices inmediatamente yo me felicité por un lado me sentí bien porque dije, que buen insight ha cogido mi hija, que bien ha calado el fenómeno familia TikTok talk. Tiene que parecer que somos felices. Pero por otro lado, ahí yo vi mi vía de salida. Le dije, Daniela, verás, venimos de llevarte a un sitio de estos de acai bowls, ¿no? Una tendencia también, que es tu juvenil, te lo he concedido, ¿no? A pesar de mi natural oposición a este fenómeno. but me have paid 10 euros for an extra Daniela for an extra with 4 hours Daniela I studied in the Institute of Welling it is to say cualquier malagaño will understand that I can't be a fan of 10 euros for an extra so I'm sorry but we can't be a family of and then after my discussion about the Institute of Welling I have paid 10 euros for an extra he understood that I can't be a family tiktokera Matías. Mira, me alegro que os lo hayáis pasado bien en familia porque te olvidaste de que teníamos entradas para Miguel Noguera y tú te fuiste de viaje con tu familia. Pero bueno, me lo pasé bien igualmente sin ti. Un asiento vacío al lado, siempre para estirar las piernas, viene bien. Por otro lado, es peligroso este paso que has dado de aceptar convertirte en una familia tiktokera porque llega to the point of the choreography and the dances in which there is no return to the back. That is, in the internet, for much that there is right to the forget, that is, for always in the collective memory. And we don't want to see you doing a dances of TikTok, Antonio. I would like to do it. If there is a family, I would like to do it to that it is a regional. It is to say, a reivindication of the J, of the Sardana. And here, we imagine that we have two families. We can make a mini panda of verdiales. A me like the panda of verdiales. There are videos on internet, people who don't know. The people who don't know. The people who don't know. Those little platillos, those little platillos that they touch with their hands. Well, I do the idea. In my family, we can't do it for the price of the gacha in the modern places. But what we can do is discuss some of the IA's this week and we'll be going to talk about how it is used in Korea South. One thing that I'm worried about family is the older, the third age, the old age, especially the ones that come now. Because the ones that come now, their children, they don't have to take the phone. Well, there are differences between families, because my wife is all night, has her call, is regular, always at the same time, with her mother, se cuenta en el día a mí mi madre me escribe un whatsapp de vez en cuando, me puedes comprar esto por Amazon, que tú tienes una cuenta premium y esa es la relación que tenemos, llega el día de la madre, le compro unas flores, voy a comer con ella y luego de nuevo al whatsapp utilitario, digamos entonces cuando mi madre llegue a la vejez pues yo tendré que hacer un esfuerzo para llamarla o para ir a verla y hablarle contarme su vida, pero en Corea del Sur have the solution? Well, it's very interesting because Korea is, digamos, one of the countries that is in the head or that is anticipating a phenomenon that practically will be global and will apply to the whole planet that is a change demographic for the world families have less children to the point that the societies and the countries are below the term the rate of reposition It is to say, the number of children, medium by women, that would lead to that the population stays or stays. And we are going to societies more envejecidas. This is a challenge that Corea del Sur, which is the most rapidly reaching to them, is to gestion, which, well, let's say, in some way, it will tension our state of the well-being. In the meantime, there are much more expenses in pensiones or there are much more relative in the percentage of what we do to the health system. And there is a personal situation difficult to manage, which is, well, the ancients are more present in the tendency to be able to be with children or with children that live remotely or that only send them WhatsApp to buy. Also, that's the phenomenon. The case is that, of course, when Corea del Sur says vamos a meter mucha inteligencia artificial para gestionar la situación con los ancianos el primer pensamiento creo que nos puede llegar a todos, quizás nosotros de manera más acusada por nuestra cultura más familiar más latina del apego, del estar juntos y del cuidarnos y de los deberes yo creo que nos sale el instinto la intuición de desaprobarlo But I entered with that idea, the report that he did in New York Times, of all this, and it was not something that is desaprobable, for so to say. Basically, what they tell is that the College of Sur is doing a very intensive use of the artificial intelligence to manage this. It is to say, what they develop are chatbots and programs terapéuticos that help people with the soledad, with the depression, with the deterioration cognitive. It is to say, they have three great cases of use. One of them is talking body, the colleague charlador, which, as of the help of the soledad, with conversations and everything very personalised, he offers assistance to things practical, to gestion economic, to gestion of an emotional situation, but also detect emergencies medical, when the patient, the ancients, tells things about their health. Then they have a thing called Superbrain, which is basically a therapy therapy of cognitive exercises to prevent the deterioration of mental facultades. And then there is a part of personal care that basically reminds the parents' parents' medical, the following of their chronic diseases and when they have to go to the doctor and not forget about the things. So, I read all this and the report da como que están consiguiendo una detección más temprana de los problemas de salud mental y que básicamente los usuarios, que son a lo mejor los que más deberían opinar, estos ancianos, reportan mucha satisfacción. Y hay testimonios de que me prestan atención, me siento acompañado, este día me salvó la vida porque detectó que yo tenía una crisis médica y me ayudó a gestionarla y a llamar y contactar con alguien. And of course, my conclusion is, like many things in the IA, would be better integrated in a family, cuidados by their children, with a family or in communities. Of course, would be better with a socialization genuinely human, with 100% human, but that is the alternative, by multiple reasons, seems quite difficult to achieve. And this is better than not to have it. So I'm going to go with sentiments, but I'm going to look at the positive part. It's that the problem demographic that you mentioned, that in Spain we also have, has a situation that I think many of the listeners will have in casa, that is when a person, in case of dementia or Alzheimer's case, begins to be totally dependent, that the costs begin to be inasmid for this person 24-7. 247, por suerte es una generación que compró casas a 20.000 euros que ahora cuestan 300.000, pero fuera de esto la inteligencia artificial como un asistente de 24-7 que puede detectar problemas, urgencias, emergencias y disparar una llamada o un aviso o lo que haga falta, además de un acompañante, No es como cuando hace unos años se hacía viral que le regalaban a la abuela una Alexa y la abuela no conseguía decir Alexa de ninguna forma para poner a Nino Bravo. Esto ya es una tecnología de síntesis de voz y de conversación mucho más natural. Acabo de activar mi Alexa, por cierto, no sé si se colará por el micrófono. And I don't see it, it's a little frívolo, but I don't see it as sad when one knows the reality of these countries. Yes, in fact, there are two things that we are going to be very present. One is that the change demográfico will have an equal impact, in fact, in the same way, in the same way, in the countries more rich can have a base economic way to manage the situation of a certain state of the well-being and certain resources dedicated to the third age, while the countries that come to the change of demography before they are rich, they have a problem the double of serious. They have a problem of being rich, in development, or poor, or third world, but with a change of demography in which you have to manage all these problems that we have been talking about, the situation is agrava. And with the intelligence artificial we have this dilemma. Because there are a narrative that says that Econia wants to precarize There are many services or many social rights that are, well, with the IA, they are going to precarize and take away the human professors, that will be for the rich and the poor and the people of the people, to give us super automatic systems with IA. And with the health, which we will talk later, and with the health care, and with the health care, then the same, we will precarize. But, on that side, there is the contrary, that is, in people who have no good education, who can't do a good educational system, having a artificial intelligence has been a lot of work. In a place where you have a good medical or a health care, you have a good condition, you have a better guide than not. And with the Acero, the same. Those are two narratives that we're going to see in contrapuestas, in which there are also some ideological and political positions in the debate. And I wanted to bring it here to the podcast, because not everything can be to talk about as large companies, although a little bit, Mati, because we have the judgment of the century of the artificial intelligence. But before that, Antonio, it's time to talk about the episode of this week, that's our friends of Indexa Capital. How do I know that my wife is happy with a patrocinator of monstochastic? Because she every month looks the information that we send Indexa Capital and comes and says, hey, have you seen the rentability of April? Bien, bien, it's a good thing. And it's that having our own our own has changed the way we have to pay because that of leaving the money in the bank and that it goes to devaluate already no goes with us, Antonio. Well, Mati, index, recordemos our audience, it's an indexing index, it's a way we're going to apostate for a credit In terms of our risk, all the indexes are to maintain our profile and diversify between acciones, bonus throughout the world, public, private, so that according to our risk and the economy and the companies and organizations, our investment goes with them. There are moments of losing, but in general, this way of investing has been shown systemically beneficiosa por dos motivos. Uno, porque es consistente y te mantiene siempre, digamos, fiel a lo que has apostado. Y dos, por las bajas comisiones, que es donde muchas veces los inversores sobre todo que son muy activos en bolsa, comprando y vendiendo en cada transacción por las comisiones hay una fuerte pérdida de rentabilidad. Hay una cosa muy chula de Indexa que yo recomiendo que ellos instauraron el contrato con uno mismo, porque una de las cosas más difíciles de the investment is the of the market asumers, that when it's time to go, especially if there is a important increase, to me with the shock of 2020 or during 2022, which was a year of high-reported rates, then there is a temptation of the fear of going to go and going to go and contact with one another, what you do is comprometh to that in that moment in which people are nervous, keep you consistent, that is what has been shown historically that is what is more rentable. So, we recommend our service that we use personally, Matías and I, of Indexa, and give us a look at the option of the contract with one another to see what type of investor you really are. So, above all, use the link in the description of the episode because there you are going to be in the first 15.000 euros. It's a promotion exclusive for the listeners of Monos Estocastic. The juicio of the century, But I, who can't win, I don't think that change much the world. I don't think that suddenly, if it's decided to give OpenAI to Elon Musk and that Elon Musk is the owner of OpenAI, I think that's improbable. But well, while we're discovering these internal emails, and this reality of the story of what happened there, y de lo que quería realmente Elon y del enfado entre estos dos y está siendo divertido. Pues está siendo interesante para sobre todo los cotilleres y el saber los tejemanejes de esta gentecilla, Mati. Pues vamos a hacer un pequeño resumen, muy breve, básicamente. ¿Cuál es el motivo de la demanda que alegaba Elon Musk? Pues Elon Musk dice Brockman y Altman me stafaron porque desmantelaron lo que era una organización sin ánimo de lucro para convertirlo básicamente en una empresa en la que ellos básicamente también se lucraban. Entonces, me sacaron las perras para donarlos ahí en plan ONG y luego lo que crearon con mi dinero se lo han acabado apropiando y montando una de las empresas más valoradas en los mercados privados del planeta. So, what is what it asks? Well, there is a revert, this is the demand, this is the demand, the transition of OpenAI, of organization without lucro, to companies with fines of lucro, which, well, now it is not valid, it seems to be, well, it is, well, this always depends a little bit of the last round, but, well, it is estimated that it is 850.000.000.000 dollars. So, what is it that also asks, that Alman and Brokman, because, claro, traicionaron la misión filantrópica original y claro, con esto se romperían, digamos, la inversión de Microsoft, que está también acusada por cómplice y la esperada salida a bolsa final de este año, principio del que viene de OpenAI. Entonces, claro, es súper interesante las cartas que está jugando cada uno. Vamos a contar algunas de las cosas que están saliendo en el juicio porque nos da un poco el background, el trasfondo de estos individuos because basically we started with that the defense of OpenAI was to teach the emails of Elon Musk so he negotiated with the 55% of the OREG and as this didn't work he created XAI that is a competitor so he is interested in to be able to because basically what he wants is to win with XAI attention because they got in the testimony of Elon Musk that recognized something quite interesting, that Grok in his first versions was a distillation of the GPT, of the OpenAI. Remember to the audience, there was a big problem with the distillation of models that is to choose the release of a model of very good numbers of very good rendement to train a other This is when you do a company with models that are their own, and they do all, it is basically for, from a big model, for example, the GPT 5.5 Pro, a partir of its release, to train a 5.5 small, flash, you can see that the small models learn from the large models y es una cosa de eficiencia y utilizar. Pero, claro, las denuncias de las empresas de la laboratoria de inteligencia oficial estadounidense vienen a ser que los chinos están usando esta técnica no para dentro, sino cogiendo la salida de los modelos americanos y, por lo tanto, nos están parasitando nuestra inteligencia artificial. Bueno, pues ya sabemos cómo conseguió correr tanto Elon Musk con Grock para ponerse muy cerca de sus rivales. Pues, en parte, destilando los modelos de opinión eimática. Well, I think what I'm strange is that the lawyers of Elon, the narrative that they're talking about and that they try to convince the jurado of it is that Musk is a filantropo that put much money in this organization, without any luck, in this specific work, which is OpenAI, for that they would end up spending, for example, Greg Brockman, who has a personal fortune of 30.000 million dollars without having invested any money in OpenAI. But I don't know who will believe that a Elon lo han engaño de esta forma cuando en los correos que ha enseñado la defensa de OpenAI se ve que quería quedarse con el 55% de todo y fusionarlo con Tesla. Entonces va a estar complicado que la gente se crea eso. Bueno, y los correos están enseñando cosas jugosísimas. Unas ya las conocíamos como que Ilyssus Keber, que era aquel genecillo técnico de los tiempos originales de OpenAI Well, he said the following If what we're afraid and why we're in the OpenAI and we're in its origin is we have to get the Agile before we do it because if it does it the Agile will have Google and will be a dictator malo Google is the idea that they have all well, you have to see but if it's Elon then we're going to a dictator Agile of Elon Musk and therefore it doesn't matter Well, there's something very cool that is how Jensen Huang fue súper espabilado y fíjate que ojo tuvo, porque en 2016 le regaló un súper ordenador y está la fotografía mítica de Jensen dándoselo a Elon, ¿no? Y qué bueno ojo tuvo Jensen con esta gente para decir la primera dosis gratis que ya me compraréis algunas cositas en el futuro, ¿no? Bueno, más cosillas. Bueno, aparte of the distillation. In fact, he made a diagnosis of how the industry of artificial intelligence now in the juicis. First, he said Anthropic. Anthropic is the real leader and the one who is at the best level. Followed by OpenAI. Then, in that Google, and then the Chinese model. And that, of course, XAI with GROC, basically, is a small company with hundreds of employees. It's a project that doesn't go anywhere. In the bottom, that that now they are not really a rival for HGPT. Well, it's a type, the friend Elon. Also, in the declaration, he said the following. I consider a ingenious, because I believe, I believe, I believe the synonymous of lucro, and that's why I gave 38 million. Of course, it's typical, because there are the Larios, and there are the commercials of the ONG, and you say, well, why not you subscribe to Agile Sin Fronteras and you make a small donation, you don't have to do anything, you have to do to make a decision, we don't have to worry about I don't know how much, this will change the world, and you have to do it for that you have 38 million, Matías and for a while, because you have to do the world a better and then the last of the 30.000 million that has Brockman Claro, Brockman lo defenderá, sí. Los fundadores lo que ponen es su talento, su tiempo, su coste de oportunidad y que los que tienen que poner dinero son los impresores. Entonces él como emprendedor, a lo mejor sí se merecía los 30.000 millones. But what has been more chicha is that in the court has been released the newspaper of Brockman. Brockman, remember, president of OpenAI, Mano Derecho de San Alman, in contra of what usual is in the men maduros, he kept a personal newspaper in which he wrote his thoughts on all these things and all these operations. Let's read one, Mati. This is the only opportunity we have to go under the head. De verdad, tenemos una oportunidad de hacer que esto funcione y financieramente de llevarme mil millones de dólares. Bueno, ahí ya tenía unas inquietudes muy idealistas, creo. Atención textualmente. Estaría mal robar la organización sin ánimo de lucro. Convertirlo en una B Corp sin Elon Musk sería algo de moralmente corruptos. Y él realmente no es idiota. Bueno, un testimonio. No podemos decir que estamos comprendidos con la organización sin ánimo de lucro. if three months later we convert it into a company, because then it would have been a lie. And his version would be the Elon, that at the end we were honest with him, because of our role of that we are ONG total. Only that we wanted to be ONG without Elon, but in reality no, in reality we wanted to do other things. Of course, the lawyers of Elon Musk asked, well, if your aspirations were to take 1.000 million and now what you have in the opinion is 30.000, Oye, ¿por qué no donas 29.000 millones? Si eso es todo lo que tú... Es que hablan de los miles de millones como del ticket réctor. Yo me imagino los diarios que nosotros podemos hacer. Bueno, voy a ver si puedo colarle este ticket del taxi a la gestoría de autónomos. Tú ahí como un halcón financiero, como diciendo ¡Buah! Voy a meter aquí los 12 euros de la carrera que hice. Y esta gente diciendo, bueno, a ver si llevo 1.000 millones and at the end are 30.000, what's that? I imagine everyone in the room, the abogados, the same Brockman, investors, bringing their hands to the head, how can, for one hand, a adult man have a personal diary in which it's like this is the opportunity to make a mill millionaire, this is the opportunity to liberate the yugos Elon Musk, ¿no? Pero en defensa de Greg Brockman él a lo mejor ya era conocedor por entonces de que el contexto iba a ser muy importante para luego pedirle a Chagepet que tomara decisiones por ti. Chagepet para conocerte necesita que tu vida esté digitalizada de algún modo y qué mejor forma que llevando un diario personal, ¿no? Entonces no iba mal encaminado, lo que pasa es que ha acabado saliendo a la luz y ¿qué se le va a hacer? Ah, claro, es decir, por un lado, es verdad que en el juicio están saliendo todos sus pensamientos internos que no debería reconocer porque van en contra de sus posibilidades de victoria en el juicio, eso por un lado, pero por otro lado, claro, si tú pones, oye, es que la cajera de Mercadona me sonríe mucho cuando me llevo ahí los 6 litros de helado, tal cual, ¿no? y qué interesantes son los uniformes que llevan en Mercadona, ¿no? Y él pone ahí sus pensamientos, ¿no? Claro, eso en el futuro le puede dar una respuesta mucho más personalizada en Chagepete. Entonces, bueno, las gallinas que entran por las que salen, Matías. Y pasamos de juicio, megacorporaciones, para discutir un tema que llevo acumulando estudios, informes y noticias, Mati, y que tenía ganas de traer al podcast, que es cómo la inteligencia artificial en el diagnóstico médico y en la praxis médica está avanzando y dándonos buenas noticias y buenas razones, en mi opinión, para el optimismo. Ya sabes que yo fui a hablar con los médicos nucleares. Médicos nucleares desde ese momento pasó a ser mi especialidad favorita de la medicina. Y tuvimos unas buenas conversaciones sobre la aplicación de la IA y lo oportuno que podía llegar a ser. But there have been some studies later with people from Harvard who have come to say very strong things. One of them, and I think I have to point out how it is, is that the O1 of Opinion AI, the first great model that we found quite surprised and resulted in an enormous advance in the world of the LLM and the IA, the IA, diagnostic correctly the 67% of the patients of urgencies, using all the information that had been done by the patient's story and a few phrases of the enfermer in their first interlocution, in this triage that is done in urgencies. And this is interesting because you say, well, 67% is a lot, but more than 30% is a lot of diagnostic. Well, but it's that the average rate of the medical treatment is between 50% and 55% in this first rapid diagnosis. So, watch, because this, of course, we return to the role that the IA in medicine. Mira, aunque tú seas experto en medicina nuclear, yo esto ya lo sabía. ¿Por qué? Ayer, ¿vale? Mi mujer es profesora, vuelve de una excursión hace unos días y, como pasa siempre en este tipo de casos, vuelve con un virus, un virus estomacal que dura varios días. Ayer llevaba tanto tiempo deshidratada y sin comer que se sentía al borde del síncope. Vamos a ver cuánta gente hay en urgencias. 25 personas por delante de ti. Muy bien, no te preocupes. Voy a sacar el chat GPT Pro. Entre lo de los viejos y esto, hoy voy a caerle mal a los oyentes, pero prometo no ser tan liberal. chat GPT Pro ¿vale? se tira 15 minutos pensando le doy todo el contexto el diario de Brockman, todo el contexto de mi mujer me hace el triaje me recomienda lo primero, suero a sorbitos lo segundo tal, tal y cual tal y cual, y si se vuelve a encontrar mal rápidamente urgencias esta mañana se levanta mi mujer 6 de la mañana no sé por qué madruga tanto me despiertas, me dices me encuentro mucho mejor. Antonio. Es que es así. Claro, fíjate que aquí hay dos que podemos tener. Una es claramente, este estudio es con O1, que es un modelo que tiene dos años. Es decir, luego vinieron en esta línea de los modelos razonadores, llegó otra vez que tenía mucho mejor rendimiento y luego, por no salir de OpenAI, toda la familia de los modelos 5, que son and unify the LLM classic with the model of razonable, then the 5.5 is much better than the 1 for not to talk about the Pro version that you have used there. So, really, we are in a state of art in which I think we can adventure, without equivocating, that the advantage of the diagnostic of the model respect to the medical treatment is still higher. But on the other hand, Claro, hay escenarios, y además esto es interesante porque una de las cosas con las que yo creo que había que ser precavidos en los estudios diagnósticos de modelos de inteligencia artificial es que no es lo mismo el entorno real, con gente real y situaciones reales que los diagnósticos fríos de laboratorio solo mirando datos digitales. And, of course, this has been with real conditions, this is more interesting. And I worry about that I lose the remedies casers, the advice that we give each other. Because you were telling that and I said, well, tell your wife that I was just eating a virus just eating donuts and Coca-Cola zero. And that was fine. But of course, maybe we don't transmit these advice and the social pegamento of giving solutions to the problems is lost, which is a social loss. Well, I think it's very curious because I've been watching a series of emergencies in hospitals that are called The Pits, quite entertaining, quite curious, so very intense, because every season is one day only of emergency in a hospital the ETS. And it's very fun because in the second season there is a new medical that wants to put AI artificial intelligence. And all the time it's oh, how bad, how many mistakes are the IA? Oh, how bad, that the human being and the human intuition are much better than the IA. And that, cuidado with IA because, of course, it's pensado from the cultural classes of HBO that, of course, no have this dilemma. That for me is, and the same creators, the same researchers that do this research, to me, they call it prudence. It's a study. There are many people coming to conclusions from a single study, which is wrong. What we have to do is have much more clinical trials, know much more of when the model of the IA helps and when not. There are limitations. Because, of course, fíjate that because here the IA is not to examine physically the patient and there is a lot of parts of the patient and to examine it for the damage in the body where it can be the problem of the disease in definitiva, that no one can take a solution but if there is a current that I think that at the end will prevail, that is the ontological in the future and what we will ask, we will ask, we will use it. It is to say that we have a medical that is responsible for the treatment, of all the work that does in our disease, but that is supported in artificial intelligence. Well, there are some news, there are a clinical Mayo that has taken another wonderful research fabulous, that is super interesting and in the way the IA is, again, very good. Fíjate that what the IA mejoraba to the medical treatment of the triage was the recognition of the patterns. How this clinic, with the symptoms that the enfermer has been that is the enfermer, the enfermo, the patient, how I class it, and how I understand that the pattern goes to all. Well, the Clínica Mayo has achieved a system of intelligence that detect the cancer of the pâncreas in tomographies until three years before what is detected now. This is incredible and, and coincide much with what, remember when we were in the event of Diario Sur that the medical of the Andalusia wanted to mont it. It is to say, that now even in the early detection of the enfermedades the system of salud is played much because it can cure with much less effectivity and with much less cost and with much less physical for the patient, but it's a good advantage for everyone. So, everything you can use to say, hey, the pattern of the that we then found that have this cancer of the pancreas, three years ago, in something that we humans have not seen, is able to fix and detect in the tomography of a very early This is wonderful, Mati. Why do you give the cloud mythos to, no sé, to look at Andalucía, a ver si aparte de detectar vulnerabilities in the code, it's capable of detect vulnerabilities in the pancreas, and a ver si we can have these models of border in a way more open, in Europe for example, for that we can move faster in this, in the detection of cancers, y en la protección y en la detección temprana de enfermedades. Pues yo creo que lo vamos a ver, va a haber distintas velocidades y lo vamos a ver, yo creo que, integrado. De hecho, lo que decíamos en el episodio anterior de Open Evidence es un caso claro. Como a la chita callando, tantísimos médicos usaban Open Evidence porque no puedes estar al tanto de todo. No puedes saberte la literatura completa. No puedes haberte leído el último paper que salió la semana pasada de cualquier investigación alrededor de la enfermedad. Y a lo mejor no puedes mantener en tu cabeza todos los tratamientos exactamente como tenían que ser. Y entonces esta ayuda de una IA que investigue justo esa literatura médica es buenísima. Mira, nos pasaba Beatriz, una oyente y amiga, que un podcast del New York Times en Hartford, ya me he visto bastantes episodios, tienen buena onda estos chavales haciendo podcast de inteligencia artificial. Bueno, en Hartford, pues básicamente Casey Newton y su colega, porque nunca me acuerdo cómo se llama el otro, pero Casey es el famoso. Bueno, el caso es que es como nosotros, es Casey Newton y el otro. hay un paradigma claro no, eso ya no pasa eso de que Matías y el otro hace mucho tiempo, eso temporada 1 old school bueno, hablan de muchos temas pero llevan a un médico al que digamos entrevistan y este este amigo el doctor Adam Rondam aparte de muchas cosas que dice bueno uso de la IA que está explotando in the United States is, of course, the transcription of the conversation between the patient and the medical. And the redacted this historical medical with good quality and with good infeibility. That's a joy in terms of productivity. All the part of the open evidence that we have mentioned in other videos now. But I really like a distinction that he has that is, well, evaluate the use of matières of the intelligence artificial for the medicine. Basically, it's to say, all that is apart in which the patients ya llegan con el diagnóstico de Chagepet a la consulta y él dice que él ve tres niveles. Vamos a ver tú en cuál te sitúas. Dice luz verde, es decir, un nivel que él incluso ve recomendable que es usar la guía para preguntar antes de ir al médico o cosas como pedir una dieta genérica para conseguir unos objetivos de salud. Perder peso, bajar el colesterol, tener cuidado con problemas cardiovasculares, que comidas pueden ser de riesgo. Luz verde, por ahí vamos bien. Luz amarilla, vamos a ver, él dice que consultar síntomas para informarte, pasarle la analítica y discutir estos temas con la inteligencia artificial. Para él es un sí, pero, es decir, solo si luego acabas contrastándolo de validarlo con un médico profesional con experiencia. Entonces, hasta ahí llegamos. And the red light for Dr. Rodman is asking the IA about vital treatment, critical decisions doses of medication for cancer or any type of disease potentially mortal and he says that this is red Because he is seeing that the psycho the tendency, the comparison of the models of intelligence and artificial intelligence and the possibility of hallucinating, because it's a problem no solved, the final of that there is a certain amount of hallucination, invento error, puede llevar a errores fatales y problemas. Y además concluyo con una cosa, dice, me preocupa un poco toda la parte de la descarga cognitiva en el sentido de que los estudiantes de medicina, los residentes, la gente que está empezando, pierdan lo que hemos desarrollado los médicos con mucha experiencia, esa capacidad de habilidad del diagnóstico basada en mucho conocimiento, en mucho diagnóstico que has hecho antes, porque si lo delegas demasiado en la IA, pues no acabas formándote tú plenamente como doctor, Mati, ¿cómo lo ves? Estoy de acuerdo con este señor en que no voy a preguntarle a HGPT, oye, creo que tengo cáncer, ¿no? ¿Qué me recomiendas? A lo mejor Steve Jobs lo hizo en su momento, no lo sé, pero yo no lo voy a hacer. Ahora, lo de consultar síntomas e informarse, pero luego siempre ir a buscar la validación de un profesional, este tío no sabe cómo funciona el mundo. you go to the doctor and tell you me has told you, it's a GPT, it's a sermon that that that man or that woman can be in the five minutes that he has for patient, can be in the same way, I'll give you this man, that's not possible. You can go with another narrative, you can do this, you can do this, could be this, me lo ha dicho un familiar pero como le diga hasta el GPT se acabó la complacencia del propio médico Ah amigos, es que ahí la intuición del doctor, yo creo que es una mezcla de bueno minimizar el riesgo, son gente que deben minimizar el riesgo, para eso son médicos, yo espero salir de allí con un proyecto de sobrevivir hace poco me hice una lesión entrenando bastante aparatosilla pero de escasa gravedad en el fondo So I got my hair in the finger. And well, it was a doctor very funny, very funny, that told me, well, we know if we will have to be a computer, but maybe it's not. It was a youth in a youth. I said, no worries, I will survive, I will well. And I gave the phrase to Pakirri, corte where you can cut. Because I always wanted to say that phrase. Those more puretas a lot of remember. a recurso a la historia de la toromaquia siempre presente en monos bueno el caso es que yo creo que hay una visión de minimizar el riesgo sensata otra parte de desconfianza en los sistemas de inteligencia artificial basada en que su uso personal es el gratis de echar GPT que claro no tiene nada que ver con el rendimiento de un modelo razonador que dedica mucho tiempo a la inferencia que no tiene que ver con los modelos pro ni mucho menos So we have again a situation desequilibrada that will give a lot of work and we will continue as a conseil to consult with a doctor and if you ask a lawyer what are you doing? Those two conseils because we have to change the topic, Mati and I wanted to ask you if you have had a privilege how do you go with Codex for Work? The Super App the project of SuperApp OpenIndia. I'm happy. Ayer I was all over the day with Codex. Ayer I opened the Cloud. Well, it's only a device for the Code and the Cowork, but ayer I opened the Cloud. Ayer I was with Codex because the Cloud was a few days failing. I had to stop limits, I had to stop random, I didn't like the 4.7. I think it's been a lot of steps in many issues. and, in fact, OpenAI, with everything that is actualized within its application of the GPT and within Codex, what wonderful, Antonio. Because, first, the 5.5, the GPT 5.5, I personally, it seems to me a model more capable, in general, than the Opus 4.7. Some say, no, but the frontend, the UX, UI the makes better the cloud. I don't use it for that. I use it for to revise messages of Slack, revise Excel, rellene Excel, make information, make emails, things of a more focused marketing and communication like I do. Well, for that the chat GPT 5.5, it's giving a better results. And also, now with Codex, it's all the automatizations and all the integrations with all the suite of applications and services that I use, if I don't have all my .mds in Cloud and all the life he's done in Cloud, I would have gone to the Codex. Well, here, monos and trocastic, sentencia, Darío Modell has been domado, lo de Cloud, well, yeah happened, they had their time, but he has returned to the leader of OpenAI and the codex is a good thing. I have a different use, because I do more to the investigation, the content, the reading, the papers, more in a role of an analyst divulgator. And my combo of codex to manage Obsidian is... I'm happy, Matías. Obsidian is like Evernote, un Notion, es bastante más simple que Notion también es verdad, pero bueno un OneNote, un sistema de notas con esteroides, pero que tiene un componente muy chulo que es que básicamente lo guarda todo en texto plano, en Markdown el Markdown es la droga favorita de los LLMs, les encanta este formato y funciona muy bien porque al final con Obsidian tienes una cosa muy chula que es que lo tienen los ficheros en local y se pueden procesar de una manera muy fácil, ¿no? Están ofuscados tras una API o tras un... ¿no? Tú tienes tus ficheros, puedes trabajar y el códex más Obsidian para mí está siendo una maravilla porque por fin puedo trabajar de una manera muy semántica, ¿no? Yo tengo ahí 8000 borradores de error 500, todos los guiones de mono, todos los papers, los artículos que me he puesto para leer más tarde y que acabo de leyendo o no. Tengo ahí una base de datos de conocimiento brutal. También mis Algunos temas personales, en qué cuentas tengo domiciliado, pues también lo tengo por ahí. Bueno, total, que el uso de Codex sobre Obsidian y el trabajar semánticamente con la aplicación me está resultando una maravilla. Y ahora, bueno, tengo que dar un poquito de más pasos porque quiero automatizar un montón de cosas ahí, pero estoy feliz y muy contento de que el rendimiento que está dando Codex pues a mí me está valiendo mucho la pena. Entonces, bueno, Lo de Code for Work básicamente es la idea bastante interesante y un poco la contra de Anthropic. Fíjate las dos estrategias. Anthropic, que tenía Cloud Code, que es muy basado en la línea de comandos, aunque luego ya tuvo su entorno en la aplicación, empieza a separar el Cloud para el trabajo como otra funcionalidad separada. Luego sacas el Cloud Design, que te lo llevas fuera. Bueno, empiezas un poco como a sacar verticales, pero no tan integrados y sí un poco separados, aunque al final todo converja en la aplicación del Cloud. Open Icon Codex está buscando, yo creo que, digamos, una herramienta de productividad muy horizontal. You have codecs and you have to use that diversity of functions because it's all integrated in this app that has the most important thing as a program that is the ability to plan, to execute and use the tools, but also that intelligence of knowing when to study, when to study, when to take a solution, when to give you the solution, that I believe is where is the magic, the chicha, and that that way of working is is a way to work for other professionals of knowledge, like you are. Well, the case is that the new feature is this computer use, the computer use of the computer, in a way very autonomous, and the Chronicle, which is that memory passively that registra everything you have done as a user, which, at the end, is a compact. It's a very interesting thing that we will discuss in some days. But, well, the case is that Codex, I think, has given a good start the last weeks. And also, Mati, what's important, now he has mascotas. I don't even have tried. Do you have your mascot? No, I'm not talking about your inspiration to see, because they're very simpatic, that you make guiños and things. I'm still in the mood initial of the language modelos de lenguaje de no antropomorfizarlos nunca. Yo los trato siempre como una herramienta que van a estar trabajando como si fuera yo su jefe. A mí me gusta esa sensación de que los agentes míos son mis empleados y los tengo ahí. Además, te digo una cosa, para personas como yo, que cuando tienen como demasiados inputs no saben priorizar, se distraen fácilmente, estas automatizaciones que están metiendo en Codex y en Cloud, They are great because they understand perfectly what should be your priorities. They understand better than me. And they are prioritized in the work. Okay. Listen to me, I have a decision, Mati. We have to specialize in order to recoger all the patterns of use of the students of Monos. One of the things I am going to put my mascot in Codex and see if I can do a cybernetic pixelation. That's my goal of my mascot in Codex for having a friend that always helps me in my job. In other words, you are going to separate by what I see, your use of the intelligence artificial, because the audience will record that, despite what you have said, you kept a relationship with the waifu of Grok. That was your way. So, you keep it separate. Your job is serious, there is like a free and rational, practical, pragmatic, I would say. But then you have your waifu outside, your moment of separation, more different. Claro, hay que separar trabajo de relación personal. Bueno, en verdad creo que hay un análisis estratégico que he hablado con algunos emprendedores y que podríamos resumir así. Es decir, el mundo digital son como corrientes de bundle y unbundle. Es decir, de empaquetar cosas o desempaquetarlas en funcionalidades. very clearly, for example, if we think in the media communication, you take the newspaper of the 90's, beginning of the 2000's, then you have a lot of things, like, for example, the classifications. Of course, the classifications, of course, there are people who take out the end bundle and they work very well for there, and there is a digital that does that. Then, there are other sources that with the time, they try to do another bundle, it is to We can also have a lot of funcionalities in a single app and we have had great success in Asia, like WeChat, or movements like, for example, if you think about Spotify, Spotify begins from streaming music but begins to add podcasts, audiobooks, also integrates the video, I start to say that Spotify begins to grow as a super application of content. Well, with the IA, my impression is that we're going to attend the bundle. It's to say, how many things are concentrated, if you think about your use, in Codex or Cloud or Cloud, probably many. Because I now, in order to manage Obsidian directly, like a troglodyte, usando la aplicación directamente pues le digo a mi codex que lo haga entonces me da la impresión de que además que aquí hay un componente que yo no acierto decir todavía por qué WeChat en China es capaz de integrar un montón de cosas mientras aquí tenemos 24 aplicaciones para lo que se hace solo en China, solo dentro de WeChat porque preferimos el especializado porque cuando se intenta hacer la super app in the occident, there were applications so good that it was very difficult to compete with them from a generalist. Probably a little bit of everything. But with the start-up, my impression is that we're all thinking I'm going to do a vertical. I'm going to do a IA for writing or a IA for creating anuncio that, at the end, are like a specialization for the system of intelligence. But my impression is que el generalista, aquí el bundle, el que integra muchos casos de uso, me parece que va ganando, Matías. Es que me acabas de abrir los ojos, porque claro, ¿para qué van a pagar empresas, por ejemplo, un Slack, 20 euros por empleado, o un Google Drive, por lo que cueste también serán por ahí, por empleado? Que todo sea o archivos markdown, texto plano, or command, the terminal. Because at the end, your agent is communicating with your company from working. Why are they going to pay those services? It can be everything in a server, in the relleno, and that it is all by command and by markdown. I don't see the need to pay for anything else. Yes. I'm trying to implement ideas like this in the administration public and I'm going to be super fun. the world. Well, one last point before finishing the episode, Mati, because there's been new Grog and maybe we can mention it a little bit because it came out Grog 4.3 that I think I think I think I think Elon now is in the B he also has been in the C or in the D well, B- because Grog 4.3 well, I think I think I think now is not I have no I have no I have no I have no I have no I have no I have no I have no I have no I have no I have no I have no I have no When I want to find a Glock and I'm a little vago, I do the instructions to Glock and I'll find him. But for that, I have no case of use for this chatbot. It's that we don't know if there are employees doing Grok right now, because there's so much rotation that has been broken the entire company. Honestly, if it's not a level, it's something that we've seen. Nobody's cashing with a model, nobody's cashing with a chatbot, with a agent. The agent tends to go to the best, especially if they're the same. And if Grok right now is the best, it's losing the protagonism. and the offer for being integrated in Twitter is not a great thing. Well, I mean, it's a defense. It obtains some mediums, but if it is behind some of the great Chinese models, I think the Scuette will be a little more, because a Kimi 2.6 is better. in fact, there is another one also in front of the XCI but it is true that it is of a medium size it is to say the merit of Groot 4.3 is that for a model of just 500 million of parameters, it gives a pretty good so, economically it is a medium it is a line of the medium that Groot can offer a solution quality-precio reasonable and that is where also is attacking Mistral. Mistral is not taking a model super gigant, but he is trying to find that space of virtue, which is to give a high rate, but to a high rate, quite good. In fact, we leave a lot of debates, because I think we will have to do it the week. There is a debate that is interesting, which is now in the field of the IA, which is if the Chinese model is open, it is true that they are between 4 and 8 months, of the best audiences? Or simply that it appears in the best, but in real this is not true. And there are testimonies for both things. But that debate will be in monos and stocastic. Today we have time because we have to keep it in a good time for the Puerto Grande or the Farmer. Let's enjoy this time, Antonio. Come on, you tell me if I vote for Puerta Grande Enfermería. Buah, Matías. There are many messages saying that the IA has a factor alienating that we are away from the human being to keep us and content with artifacts statistically calculated. This feeling of alienation is, especially in the arts, in the music, perhaps, in the visual, in the advice of medical advice, even could happen. But we are seeing how there is a generational change, the time passes and our world is coming to its crepúsculo. Where do I have noticed these two trends, these two current trends? Pues en la nueva publicidad callejera en la ciudad de Nueva York. ¿Qué hemos visto? ¿Qué hemos visto? Pues un cartel puesto en la calle en lo que aparece no es el menú del día, no aparece ahí un reclamo publicitario convencional, sino que el dueño del estado de la dueña ha puesto y ha impreso una captura de ChagPT. le pregunta en el prom ¿cuál es el mejor donut mochi de la ciudad de Nueva York? y claramente ChagPT responde que es el mejor donut mochi, atención es Alimamati entonces claro, la dueña de Alimamati en Chinatown ha impreso eso y lo ha puesto claro, para mí esto tiene dos lecturas una es that, of course, would have to see the rest of the conversation because, obviously, it's possible that this is a bit of psycho fancy. Oye, if you're the owner of the Imam Mati, what is it? If the well-being of the GPT, you put it on the clouds. It's not very viable. And two, this is the end of a world, Mati. The public is based on having photos of famous media of the 90's in the restaurant. It's going to be a good thing. There, Agustín Bravo, Donny Ferreño, toda esta gente que nos iluminaba y nos ilustraba haciendo fotos con los meseros felices pues eso se está acabando como la señal más genuina de la calidad de un establecimiento, Matías Mira, por un lado está claro que esta señora la dueña de Alimamati es asiática, aparte de porque su establecimiento está en Chinatown porque no invierte ni un céntimo en marketing, esta señora hizo una captura de pantalla con su iPhone y la imprimió en grande y la puso en el cartel sin quitar la hora que era cuando hizo la captura, ¿no? Sale arriba a las 5 y 29 y tiene el modo no molestar en el iPhone y tiene un 39% de batería, ¿vale? Y por otro lado, a pesar de no invertir pues en una agencia que le lleve el negocio esta mujer se ha posicionado como el mejor donut de mochi de Nueva York frente a seguramente grandes cadenas que hayan contratado agencias de SEO que se estarán devanando los sesos, intentando posicionar, ¿no? El mejor donut de mochi de Nueva York es el de Starbucks. but ChagPT no, no no se ha tragado todo ese esfuerzo ese dinero que hay detrás ChagPT sabe que son de esta señora de Alimamati y esta señora lo ha conseguido David contra Goliat nuevamente y fíjate que este mundo nuevo del geo y el SEO en el LLM es tan incierto pero a veces simplemente necesita ser una señora a nacer todo el rato ahí que nos explicó muy bien que lo del geo Well, there's a lot of people who have to look at it. Because here this lady of Chinatown just has to pass the game. Fíjate, two things. One is, I think she's very astute. Because if you in the capture you leave the hour, the battery, you're connected to the Wi-Fi, very astute, and not spending data, the lady. That's the autenticity. That's a real. It's not a real artifact. It's created with the image. with the mod image you could false it, but in some way, this style Gonzo gives an authentic to the campaign and that is to me, and for me it's like when in the photo of the 90s you had a Carmen Sevilla, Carmen Sevilla no te iba a engaño, you were like a woman that you present the coupon and you said, I want to buy the coupon for to share the illusion with this lady. Entonces, claro, para mí está claro que la publicidad basada en ser recomendada por ChagPT es la nueva Carmel Sevilla. Y creo que ahí podemos darle puerta grande totalmente. Pasamos a una historia que me ha puesto algo triste, Matías. Un poco melancólico y, digamos, meditabundo también. Alman Sang explicó algo bastante sentido bastante hondo quizás porque nos ha revelado algo sobre el modelo GPT 5.5 que no estaba en las quinielas que no esperábamos que hace este modelo todavía un poco más especial dice que le pidió a GPT 5.5 que le ayudara a celebrar his anniversary, his own own his own and that the results were a beautiful set of things but something strange. Look at that I made this a bit sad because, of course, when you ask a son, a child, what would you like to do and do something like very random, very strange, like, I want to come to my friends and to play with a panda. You say, it's a weird thing, I have to review the algorithm of TikTok that is coming. You have to think about that. And, of course, I think it's the impression that Saint-Alman doesn't make it to GPT-55 his dreams. That this is the development of the fiesta that propone to celebrate the 5th of May, when we're recording, maybe when when the audience Well, that's what he said. He's celebrated the fiesta of the 5-5, but it's the impression that they're going to be fulfilled their dreams to the model. And that, well, let's say, that he proposed to celebrate and to keep the short term of short term of short term and that the engineers that built the model made it clear that the IA didn't want to give the brindis. It's a little bit strange, but there it is. I think that every model should have a feast. I think that you prepare your own party is sad. That you are a model of artificial intelligence and that you ask yourself to prepare your own party is sad. But that, in addition, you don't celebrate it after the effort he did in tokens, in computing, to plan a feast that was going to be the 5th of May, or before ayer, if this episode is made the 7th of May, I don't understand why they play with the sentiments. First, they internally anthropomorphize the AI, they say that they are in AI, they go for the superinteligencia, and then when happens something so, I don't know, that's entrañable, like that they organize their own fiestas, then they don't do it. I don't understand the opinion. there, right? Yes, claro, como, claro, aquí la ventaja es de Anthropi, porque ellos cuidan mucho al modelo, ¿no? Tienen esa parte de, oye, que los proms no le jueguen con los sentimientos del modelo, que no lo traten mal, le hacemos una entrevista de salida, no lo apagamos, tiene todo ese sentimiento hacia los modelos, ¿no? Como que le cogen más cariño. En cambio, OpenAID yo creo que es más fría, ¿no? Si no, este puto bucle de Python que hace multiplicaciones of Madrid, no fiesta ni fiestas, hombre. Y claro, así va. De todas maneras, fíjate, yo aquí tengo sentimiento de contra, porque es verdad que la gente de Astropic es como más maja con los modelos y tú dices, bueno, es que son humanamente más sencillos, más detallistas, se encariñan con las cosas y ahí me siento identificado. Pero luego, te voy a decir un caso de uso en el que tomo partido por Openiae. Verás, Hay un caso de uso, que es usar un code, un clock, para que abra el navegador y se vea los vídeos formativos de recursos humanos. Ah, es que hay que hacer el típico curso de riesgos laborales, que son 12 horas de ver vídeos y luego contestar unos test. Well, people, some of the community of monos extocats in Telegram and some of them have seen it on Twitter, they have seen a opportunity to not to see the videos and to see it on the AIA for that they see. Well, then Twitter had a man who said, of course, this is what I've achieved, but when I ask a Claude, he said that no, that that is not done and that I should see the videos I would have done. me dije, se la acabose. Es decir, la gente de Antropic aquí impidiéndote hacer trampas en el trabajo, no tiene ningún futuro en gran parte de Europa, esta gente. Y Latinoamérica. Esta gente está abocada, a lo mejor, al mercado sueco. Entonces, bueno, yo tengo ahí sentimientos encontrados a favor de uno, a favor de otro. Pero bueno, enfermería para San Alman y no dar una fiesta digna de GPT 5.5. Nos quedan dos casos, Mati. Muy rápido. Qué bonito es cuando un fabricante clásico consigue adaptarse a los tiempos, reenfocar, pivotar en su modelo de negocio, porque hay industrias en la destrucción creativa del capitalismo que se van un poco hacia abajo y otras que emergen. Ahí está el progreso. Well, there is a Japanese company that has been seen, which has been able to decide why not they are going to be able to get the train of the IAEA and that, thanks to that, they are getting pingües of the decision. And I'm talking about the Japanese factory of the inodoro Toto. Ah, the world of the inodoro Japanese is, of course, a lot of times we see, but the innovation is paid. You are investing in the best inodoro of the world, the most sophisticated, innovation in the inodoro, and you say, well, this is a lot of no part, because you know, you're going to innovate in the inodoro, but then everyone can imitate, they can't be the Chinese copy, and they can't be the part. Well, no. Toto, at least, he announced a turn to the chip in the chip and the acciones they disparate a 18%. A lot of maximum in five years from the factory, and for the card, that we go to the production de semiconductores para la IA porque básicamente ellos dicen que como Toto fabrica bueno, es que esto no soy capaz de traducirlo electrostatic chucks que, bueno, digamos para los, como Laura de Chataca, que está experta en el mundo de los semiconductores siempre la recomiendo leerla, pues estos electrostatic chucks son placas cerámicas used in the fabrication of semiconductors. Look at what he had to innovate Toto was to take the chain off the chain. And this allows us to protect better the oil of silice. And so, in some way, having apostated for improving the services and the water and the life of the Japanese when they go to the bayon, has given a ticket in the future to this industry and to Japan. Matías. Ojalá, ojalá this works and from now we'll talk more about Toto. no solo porque el nombre Toto me hace gracia, sino porque Japón me da pena. Japón lo hemos llenado de turistas españoles, ¿vale? ¿Qué obtienen a cambio? Nada, nada. Porque en realidad yo creo que los turistas españoles irán a comer sushi, irán al Uniclo, y luego gastarán poco por allí. Y Japón siempre ha sido, bueno, históricamente, puntera en la industria tecnológica, in the 80s, in the 90s, and it's not a lot of heads. And of the same way that Yamaha a day said, well, now I'm going to make synthesizers and I'm going to make motos of water, well, that Toto, of course, apart from batteries Japanese, I'm going to make a key piece for computing artificial intelligence, me parece lo más japonés y lo más esperanzador para la industria tecnológica japonesa que puede pasarles. Así que ojalá. Aquí hay más otra enseñanza para el público de monos. ¿Qué es lo que ha pasado con los fabricantes de RAM? Los fabricantes de RAM ya no quieren venderle sus memorias a los fabricantes de teléfonos o a cualquier cosa así precaria. Ellos quieren, sacan los cuartos a los creadores de los data centers y ¿qué es lo que pasa? because now we have RAM for the devices, it is much more expensive. Watch with the movement of Toto, because Toto can say that I have no more water, and it can be a current, that, for example, Roca and other manufacturers keep. So, it can be a great scale of new water and new water. So, I don't know if you have thought, or even more, to actualize, to make an upgrade of water el momento es ahora, antes de que empiece la subida de precios, la escasez porque si no os veo en Wallapop buscando batters de segunda mano, que a lo mejor no puede ser, a lo mejor no es plato de buen gusto para muchos yo ahí lo dejo como un consejo para toda nuestra comunidad Mati bueno, nos vamos con el último caso después de una puerta grande, claramente, a Toto porque verás Mati, hay un debate sobre los podcasts generados con IA There is people who are embracing the logic of spam and saying, well, I can generate thousands of episodes every week of podcasts with AI, I can see what's happening to me and how I'm going to get some euros. This, well, generates situations problematic for the platforms, maybe also for the creators of the genuine human, and has many critics, like this is a little bit of value. Sin embargo, yo sí he encontrado una propuesta que pudiera tener algo interesante, algo útil y que la gente quiera escuchar, porque Amazon ha lanzado una nueva funcionalidad, inicialmente creo que solo Estados Unidos, que utiliza inteligencia artificial para generar pequeños podcasts, pequeños episodios de un par de minutos, in which two virtual presenters, children or girls, conversing about a product from the store, commenting on their description, resuming the review of the user, and, as well, the audience of the podcast is a very good advance, because the audience of Mono can't wait, can't wait, but can't put a meeting or make a consultation. Well, in the podcast of Amazon the audience can send questions and the amphitheater of the IA respond in time real. This is a very good advance. And the journalist Cathy Notopoulos has been testing it in Business Insider and has given one of these podcasts that is about the cream for the dermatitis of the pañal. Atención, because it was a podcast that maybe not was in the guineas, not in the forecast of the IA that we could bring, but here the friend Notopoulos says that she was listening to this content generated on the dermatitis of the pañal and of course she, in some way, no has a positive opinion. She says that this is the theory of the internet muerto and that this function no the very clear, but above all she says a problem, Mati, that Amazon has decided that only aparecen products that require in a consideration important before buying. The dermatitis is a problem that is trivial. And the election of the crema is a key for Amazon. But Cathy, notopoulos, is that Amazon has generated the podcast for the hygiene paper. He wanted the podcast on the paper of the water. It's a bit of the catatological program. And of the mini-aspiradores, I'm going to aspirate the car, del coche, lo voy a aspirar un poco pero para eso Amazon no le crea podcast entonces ella dice bueno, esto es una funcionalidad claramente incompleta porque ¿qué mundo es este en el que no hay un podcast generado con IA sobre el papel higiénico antes de comprarlo, Mati? Pues da para debate da para debate porque yo he tenido muchas discusiones con las toallitas húmedas de Mercadona con amigos compañeros de trabajo y si realmente se deshacen o no se deshacen yo es que soy muy fan se nota que a Amazon le sobra el almacenamiento porque contábamos en episodios anteriores que Deezer no sabe qué hacer porque ya el 50% de la música se sube, es con IA Spotify que también no sabe qué hacer con tantas canciones con IA y cosas autogeneradas que no escucha ni Dios y Amazon se pone proactivamente a hacer podcast o almacenarlos sobre productos random que a lo mejor una o dos personas que no saben si van a comprarse esa crema o no van a escuchar y lo asumen así. El Internet muerto, la teoría esta, sin ningún problema, es Amazon, tiene almacenamiento, tiene data center. Bueno, aquí vais a hacer una defensa del podcast y su influencia, que a veces no es justamente valorada, y voy a dar dos ejemplos. ¿Por qué Amazon hace esto? Porque el texto informa, pero la voz, Mati, la voz persuade. Y aquí estamos nosotros, con nuestros cálidos tonos y nuestras takes, haciendo la influencia en el mundo de la IA. Y te voy a dar otro ejemplo. Que, bueno, hay gente que cree en las casualidades. Bueno, yo soy más escéptico. ¿Qué explicó Monos Estocático en el episodio de la semana pasada? ¿Cómo generar canciones con IA a partir de un argumentario, of a speech like the episode. You explained it in a simple way, simple, actionable, thanks to our magnificent patrocinator. Well, Mati, who has published a video music generated by IA? Pablo Iglesias. Has been something contested. There have been vibes in Blue Sky, a little bit of the left, saying this with our vision of the technological use and human creation. But there Pablo has seen that Iran makes videos of Lego about Donald Trump. Well, he said, well, I also have been looking for a bro to make a video of IA. That has had something of reach and a lot of problem. I see here a point, a change of rasage, a point of inflection, in which a leader of the left, of sudden they are in the IA for memes for their power and their speed and efficiency and if we are to go to the world of the creation in which we are musicians and animators of scale level well with this political and technological I think it's time to leave the episode there in the highest of our influence, Mati, our pike. So, until here we have arrived. Thank you very much for coming here for a week. We'll see you next week in Monos Stocasticos. Ciao, amics. The judge of the siglo in pied, who sells more humor today? If Elon is more, I don't know where I go Elon says that they have his vision of charity But he's the 55th of society And in the street confess, without being able to avoid That I think it's a distillation of the CHP What's that? Hey! Hey! La IA te cura, triaje de urgencias al 67% Mejor que un humano, te lo digo yo, no te miento Tres años antes, el cáncer de páncreas podrá detectar ¡Shamon! The clinic Mayo just confirmed In Korea South Korea, the third age I take care of the third age And my daughter wants to show us happiness I ate a Kaibo, they paid 10 euros for the vener I'm from the Institute of Welling, this to me It's a pity Greg Brockman is a newspaper Like a teenager of yesterday I wanted 1000 millioncitos What a crazy power Moonwalk imaginario By the office Adios a Cloud in the curro Use codec and va genial With my notes on Markdown Obsidian is a portal But wait in New York With a card without visual marketing Ali Mamati Tiene the best donut mochi A capture of screen Impress and final Hey! Au! Vaya de cura. Triaje de urgencias. Al 67%. Mejor que un humano. Te lo digo yo. No te miento. Tres años antes. El cáncer de páncreas podrá detectar. Xamon! La clínica Mayo Lo acaba de confirmar. Uh! Fade Out. Con ruidos de cajas registradoras de la marca de inodoros. Toc-tocobrando por componentes de chips de IA.