maranosa (@maranosa) profile picture
Joined September 2010
Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)Photo by maranosa (@maranosa)
Este punto en cámara lenta... — La reacción de Medvedev. — El festejo de Alcaraz. Cine en su máxima expresión 🍿
👀 ¡MAÑANA SE DEFINE EL GRUPO JOHN MCENROE EN LAS ATP FINALS! Así están las probabilidades de victoria en Kalshi: 🇺🇸 Ben Shelton: 65% 🇪🇸 Rafael Jódar: 35% ¿Quién queres que gane?
11
175
12
2,246
214,521
Some random, unfiltered thoughts on AI and the market: The gulf between “San Francisco” and “New York” has never been wider than it is right now AI is accelerating in capability, rapidly and in a way that is structurally changing, and in many ways upending, the nature of the firm and the fundamentals of capitalism If that sounds extreme, consider that Anthropic has ~6,000 employees, Meta has 67,000, and both companies are worth ~$1.5T Talk to your average startup and it’s clear that, given the opportunity to start from scratch, nobody is building their companies “the old way.” Talk to your average Fortune 500 company and it’s clear that getting an AI agent into production at scale is a Herculean task Intelligence may not be the bottleneck, but firms built from the foundational assumption that intelligence is improving quickly in capability and price are scaling at a genuinely unbelievably velocity There’s a reason Zuck wanted to fire 60% of many of the teams at his company, in large part likely out of frustration that he can’t start from scratch like his competitors had the luxury of doing A structurally weaker Nvidia and OpenAI have minimal impact on the pace and impact of AI diffusing across the economy (“San Francisco”), but completely breaks the models built in “New York.” Hence the absolute freak-out these last few weeks. It was all priced to go perfectly and now things are not perfect But for the boots on the ground, AI will not slow down. And counterintuitively, economic weakness might accelerate diffusion. After all, tokens are much more fungible and flexible than people Where does that leave us? I think firmly in the second phase of the AI build-out. Real value is being built in the app layer (though how defensible it is…TBD). Trying to predict the next bottleneck, or whether Washington does this or that other thing wrt policy, or exactly how many data centers get built out…who can say? Certainly not me. It’s not the bet I’m trying to make I’m looking at who wins from rapidly improving intelligence available at increasingly cheaper prices, and critically, which leaders are able to move with conviction into this new operating paradigm
82
64
19
1,539
164,780
📹 El disco de oro de las ‘Voyager’: lo que Carl Sagan le quiso contar a una civilización alienígena Explore los saludos en 55 idiomas, los sonidos de la Tierra, la música y las imágenes, grabados en un ‘vinilo’ dorado que desde 1977 viaja por el espacio y que encierra también un intento por registrar cómo se siente el amor elpais.com/ciencia/2026-11-1…
6
249
12
696
22,365
"Tengo las rodillas despellejadas de mirar tanto debajo del coche", dijo Ernest Lluch en septiembre de 2000. #ETA lo mató a tiros dos meses después #TalDiaComoHoy. #Vivirsabiendoquetevanamatar #Memoriareciente ¿Me ayudas a recordarle? Si lo hacemos seguirán viviendo en nuestra #Memoria
130
2,593
81
5,580
118,526
maranosa retweeted
El 21 de noviembre de 2000 murió asesinado por la banda terrorista ETA cuando tenía 63 años Ernest Lluch, destacado intelectual, economista, profesor y político español que ejerció el cargo de ministro de Sanidad, desde donde impulsó la sanidad universal. Catedrático de Historia del Pensamiento Económico, economista y político catalán con raíces maternas en la Horcajada (Ávila). Ernest Lluch supo compatibilizar la participación en la política activa con su dedicación a la actividad académica. La lucha por las libertades y a favor de la democracia durante la dictadura de Franco le acarreó problemas tanto con el Ministerio de la Gobernación como con las autoridades del Ministerio de Educación y Ciencia. Los esfuerzos en los últimos años a favor de la pacificación del País Vasco a través del diálogo y la negociación política fueron segados trágicamente por la intolerancia y la barbarie de unos cuantos. Miembro destacado del partido de los socialistas catalanes, en 1977 fue elegido diputado para el Congreso de los Diputados por Gerona, cargo que repitió hasta 1988. Fue nombrado ministro de Sanidad en el primer gobierno de Felipe González en 1982 hasta 1986. A él se debe la Ley de Sanidad de 1985 que significó la universalidad de la atención sanitaria en España. Pero la verdadera vocación de Ernest Lluch fue la dedicación a la actividad académica, actividad, que conectaba con la defensa de sus ideas políticas. Desarrolló su actividad académica en una triple dimensión, como profesor, como investigador y como maestro que supo formar una auténtica pléyade de discípulos. Como profesor enseñó en las Universidades de Barcelona y Valencia; primero como profesor ayudante en Barcelona, después como profesor agregado en Valencia y finalmente como catedrático de universidad en la de Barcelona. Mención especial merece su cargo como rector de la Universidad Internacional Menéndez Pelayo de Santander entre 1989 y 1995. Como investigador es necesario destacar sus estudios sobre la historia del pensamiento económico español en general, pero, sobre todo, catalán, fundamentalmente en el siglo XVIII. No cabe duda de que hizo avanzar considerablemente aquella disciplina en España, fomentando la recuperación y edición de los escritos de los economistas españoles del pasado, pero sobre todo puso especial cuidado en la formación de un grupo numeroso de investigadores y profesores de Historia del Pensamiento Económico. Lluch realizó un esfuerzo considerable por destacar aquellos aspectos más positivos de los mercantilistas y prestó especial atención al estudio del cameralismo y su influencia en el austracismo, cuyas derivaciones políticas ilustrarían su defensa de una determinada descentralización política. En su búsqueda de soluciones que pudieran estar ocultas en procesos históricos, para planificar un futuro más justo en un estado de las autonomías sin victimismo, escribió que “la derrota de las libertades catalanas era paralela a la de las libertades españolas, en el sentido de la época, y en común era también el enfrentamiento con el absolutismo unitario que acabó triunfando. Nuestros austracistas deseaban que volvieran ‘Las Españas’, incluidos los programas de los hermanos comuneros”. Aunque no dejaba de lado los aspectos analíticos, su enfoque siempre tuvo un carácter más relativista, en el sentido de resaltar la importancia de las circunstancias de tiempo y lugar, esto es, los entornos políticos y sociales en que surgen las teorías, lo que le llevó a resaltar la necesidad de analizar las teorías económicas desde una óptica nacional. Como maestro puso especial atención en la formación de sus discípulos, la mayoría de ellos hoy profesores de Historia del Pensamiento Económico, catedráticos y profesores titulares que enseñan en las Universidades de Barcelona, Valencia, País Vasco, Aragón, Andalucía y Galicia. Tal vez no sea exagerado hablar de una auténtica escuela, como el propio Lluch reconocía con “una influencia teórica neorricardiana y/o neomarxiana y/o neokeynesiana con una visión intervencionista o proteccionista”. Unos meses antes de su asesinato, había publicado un artículo en el que exponía su tesis sobre la primera víctima de ETA en la que defendía que no fue ni el guardia civil de Tráfico José Pardines, ni el policía franquista Melitón Manzanas, sino la bebé Begoña Urroz Ibarrola. Vinculado a San Sebastián y miembro de la organización pacifista vasca Elkarri, Lluch destacó como defensor del diálogo y partidario de establecer puentes como instrumento para conseguir el respeto a los derechos humanos y la pacificación del País Vasco. Todo esto fue determinante para ser elegido como víctima por ETA. El 21 de noviembre de 2000 fue asesinado mediante dos disparos en la cabeza por la banda terrorista ETA cuando se encontraba en el garaje de su domicilio situado en la avenida de Chile, de Barcelona. Su cuerpo fue descubierto por un vecino una hora y media después del atentado. El asesinato fue reivindicado por el llamado Comando Barcelona de la banda terrorista de ETA. El asesinato de Lluch provocó un rechazo abrumador en España, culminando en una gran manifestación en Barcelona dos días después del suceso, con la participación de cerca de un millón de personas. “Estoy convencida de que Ernest, hasta con la persona que le mató, hubiera intentado dialogar. Ustedes que pueden, dialoguen, por favor”, exhortó la periodista Gemma Nierga durante el evento.
1
12
1
30
1,332
"Shoot them like vile dogs." Andrei Vyshinsky, Stalin's chief prosecutor, closes the trial of Nikolai Bukharin, March 1938. Bukharin, Rykov and sixteen others were shot four days later. English subtitles.
6
20
2
84
7,336
If you are European, this table (which I adapt from @Afinetheorem) should make you pause for a second. These are the top 12 companies in the world by market cap yesterday. A few observations: 1) Ten are from the U.S., one from Taiwan, one from Saudi Arabia (the Saudis are just lucky, so let’s forget about them). 2) All of them younger than me (Microsoft, the oldest, is from 1975). 3) None of them are in the financial sector. All of them are in what is called (I think imprecisely) the “real economy”: computers, chips, AI, social media. 4) Not a single European firm. Are these market caps the reflection of something real? If I look back at the last 60 minutes of my life, I have used products from Apple (the computer I used to write this post), Microsoft (I wrote the first draft of this post in Word), NVIDIA (I proofread this post with Claude, which most likely runs on NVIDIA GPUs), Google (I used Chrome to open X), Amazon (I was listening to Amazon Music while I wrote this post), SpaceX (this is X!), TSMC (the chip in my computer), Broadcom (most likely the designer of at least one of the semiconductors I am using now), Meta Platforms (I just sent a WhatsApp message to @lugaricano about how tough the task ahead of Hernández de Cos is), Tesla (I own a Tesla), and Anthropic. What about Saudi Aramco? I have probably used some imported product that uses oil sold by Saudi Aramco, but that is harder for me to determine. What do I need to say to awaken Western Europeans from their slumber? The original post by Kevin: teapaw.org/Afinetheorem/status/21…
Today in Schumpeter: the ten US companies with the highest market cap (10 of the world's top 12) were founded after 1975. 4 in the 20th century. 1 in the 2020s. 8 of Europe's 10 biggest are over 100 years old.
96
271
22
1,281
91,532
1. An Introduction to Economics, by Thomas Sargent, Nobel laureate 2011. In 2007 he gave Berkeley graduates a 335-word commencement speech: twelve lessons and nothing else. Here they are, one by one. The examples and suggested readings are mine, not his. 1) "Much of what we want isn’t feasible." We want more services, lower taxes and less debt. Not every mix is possible. Economics asks which ones are. My suggested reading: Robbins (1932), An Essay on the Nature and Significance of Economic Science.
14
174
28
752
99,025
It was nice to be in Madrid last Wednesday and meet a lot of people in the award ceremony of the Spanish Math Society (RSME) & the BBVA Foundation. I had to give a short speech and (of course) I talked about AI & Math. I share that part here: " I believe we are living through the greatest transformation in the history of mathematics, greater even than the axiomatization of the late 19th and early 20th centuries. We mathematicians, and society as a whole, must ask ourselves what the value of human understanding is now that a machine can prove theorems that humans cannot. This is an open debate. My view is that we don't do mathematics to know the answer, but to understand it. For a mathematician, being convinced that a statement is true is not enough. We need a proof because the proof allows us to understand. And that understanding is what mathematics contributes to other sciences and, ultimately, to society. Today we must stand up for that goal. A future where the human role is to understand and pass on what machines discover may sound sad; but a future where we could not even understand those discoveries would be far sadder and more dangerous. All of this forces us to rethink part of our identity. We are mathematicians, in part, because we are good at proving theorems. That has been a large part of our motivation and of our contribution to society; it has earned us the recognition of our colleagues, it has given us joys and sorrows and, for those of us here today, it has won us a prize. That source of motivation is coming to an end. But I want to focus on another great source of motivation for me: curiosity. I am a mathematician because I want to understand, and because I want to teach what I understand. My ego cares whether a theorem is mine or the AI's. My curiosity doesn't care: my curiosity wants to know, and makes me ask new questions. And that is why I believe that, in this new era of mathematics, we need less ego and more curiosity. "
Recoge el premio Vicent Caselles Gonzalo Cao Labora @GonZalocla (@EPFL) por sus contribuciones al estudio de las singularidades en los fluidos compresibles como el aire #PremiosMatemáticas
6
11
2
55
4,187
➡️ @judith_arnal realiza un análisis riguroso de la economía española y global en esta interesante entrevista de @fpasc: 🔹 Fondos europeos: Aún quedan 22.000 M€ pendientes de desembolso. La ejecución ha sido demasiado centralizada y atomizada. Mientras tanto, la inversión empresarial en 2025 seguía por debajo de niveles de 2019. 🔹 Vivienda: con un déficit de 750.000 unidades, desde 1980, el precio real ha subido el doble que la renta disponible. Intervenir rentas o estimular la demanda sin desbloquear la oferta "echa más leña al fuego". 🔹 Pensiones: su gasto aumentará en unos 4 puntos su peso en el PIB en 2050 y el debate no es si se pueden pagar, sino cuál es el coste de oportunidad sobre sanidad, educación o infraestructuras. 🔹 China vs. EE. UU.: China nos aventaja con su "shock 2.0" industrial y prepara el 3.0 (IA y robótica). En España no sustituye a EE. UU.: el capital estadounidense genera 344.000 empleos en nuestra economía y el chino, 14.300. 🔹 Productividad: ha crecido un 0,4% entre 2022 y 2026, por debajo del periodo 2014-19. Sin empresas más grandes y mejor talento, la productividad y los salarios no despegarán. Artículo completo: elmundo.es/opinion/luces-par…
4
35
3
68
4,293
It's difficult to have an adult conversation about pensions, because for most people it's built on two false premises: That pensioners are poorer: Hasn't been true in many places for 10-20 years And the biggest misconception: that pensioners are getting back what they paid for
44
433
20
2,555
84,409
maranosa retweeted
Para poner la cifra en conexto, la película más taquillera de la historia que es los Vengadores Endgame hizo 2,94 billones
BREAKING: Rockstar Games says Grand Theft Auto VI sold $3.79 billion in 24 hours, making it the biggest entertainment launch of all time. Pre-sales of the game reached 32.5 million units
6
45
1
289
18,627
Some are saying "proving the Riemann hypothesis makes no meaningful difference to anything". But it means thousands of conditional results are now also known. More than that, it shows that the prime numbers, the basic ingredients of the quantitative structure of the universe, obey a regular pattern forever. This is immensely meaningful.
For those who dont know what Reimann Hypothesis means : Prime numbers (2, 3, 5, 7, 11…) are the building blocks of all numbers, and up close they seem to pop up completely randomly. However, if you zoom way out, we actually have a smooth formula that estimates roughly how many primes exist up to any huge number and it’s just never exact. ​In 1859, a mathematician named Bernhard Riemann found a way to turn the prime numbers into a 2D map filled with special "zero" points, where each point acts like a wave that controls how far off our estimate can get. The Riemann Hypothesis is the claim that every single one of those special points lines up on one exact straight line, with zero exceptions, all the way to infinity. If that’s true, it guarantees that prime numbers never clump up or vanish in weird ways and they stay spread out as cleanly and evenly as random coin flips, forever.
116
306
32
3,190
178,515
The gap between China’s and the U.S.’s real output growth has decreased rapidly over the last 15 years. By around 2010, it was over 8% per year on average. Today it is 1.75%. To show this, I plot the real output growth for China minus that for the U.S. between 1991 and 2026 (1991–2025 comes from the WDI; 2026 is the WEO forecast as of October, so it’s pretty close to the final number). I also plot the trend to smooth out cyclical fluctuations and gauge the average gap. Nearly all the reduction in the average gap comes from China growing more slowly, not the U.S. growing faster. A world where China grows, on average, 8% faster than the U.S. is very different from a world where China, even if richer, grows just 1.75% faster. But the interesting question is what will happen to this gap in the near future. If China follows the same evolution as Japan and South Korea when they were at a similar stage in the growth gap (and China has replicated both countries’ growth-gap patterns quite well so far), the gap will close by around 2036–2040 and become negative afterward. Of course, China may break the pattern and behave differently than Japan and South Korea. I am not particularly optimistic that this will be the case. I have a paper: sas.upenn.edu/~jesusfv/Neocl… where I argue the date might be even sooner, 2034, because China’s demographics and productivity growth are worse than Japan’s and South Korea’s at similar moments in their development. But a few years up or down do not make much difference. I will try to find time to show some additional parts of the argument over the next few days, but not tomorrow, when I will be giving a seminar at the Columbia School of International and Public Affairs. I will present some cool new shipping data on the effects of closing the Strait of Hormuz.
19
78
6
394
26,831
This slide from the Anthropic IPO roadshow was a big factor in my decision to leave. It’s an insane plot and I’m surprised they made it public. It shows the compute efficiency of pretraining over time. The y-axis is a **log scale**: they are seeing continued superexponential algorithmic progress. Mythos was not a one-time deviation, but a point on a superexponential. They could already retrain Mythos with a fraction of the compute originally needed. It doesn’t strictly imply a software-only intelligence explosion- they have been spending more on R+D. But algorithmic improvements do not require ridiculous creative genius and are on the verge of being automated. You do the math.
399
549
296
6,305
1,740,007
maranosa retweeted
Replying to @sciam
15
75
1
988
105,915
跟几个行业里的高管聊了聊,我听到几个让人坐不住的消息: 1. 老黄再不回来,NVDA 可能要迎来一波大辞职潮。 数据不会说谎:英伟达员工2026财年的流失率已从2.5%升至3.7%。78%的员工身家超百万,近一半超2500万美元——想退休随时能退。他们还在干,就是为了老黄。 现在临时CEO(原CFO)上任,大家“再撑一下”的借口没了。管理层承诺加薪、增发股票,但钱解决不了信仰问题。更麻烦的是,董事会已启动新CEO遴选,头号目标是Arm CEO雷内·哈斯——他在英伟达干过7年,算“老熟人”,但空降能补上老黄的窟窿吗? 2. 市场对NVDA未来增长的信心在松动——老黄本身就是护城河。 媒体很少提:老黄特别擅长自上而下的架构设计,关键决策靠他“玄学级”的判断力。CUDA、网络业务、内部团队吵架,最后拍板的永远是老黄。这些决定没法写进流程手册,但每次都决定了Nvidia+整个AI生态往哪儿走。 然后供应链跟着走、资金跟着走、协议跟着走、信心也跟着走。 空降CEO或提拔内部高管,都补不上这个窟窿——台湾代工厂跟老黄攒了十几年的信任,不是换人就能无缝衔接的。老黄一次次押上改变行业的重注,硬是把Nvidia干到5万亿市值。 3. 供应链已经开始松动了——这才是最要命的。 台积电、SK海力士这些伙伴,态度都在变。以前老黄在,都是抢着配合;现在要求英伟达先付预付款才肯排产。 数据更吓人:英伟达的供应与产能承诺达2790亿美元,比上季度暴增1600亿。这些钱是用来锁HBM内存和台积电先进封装产能的。信任没了,合作就要付出真金白银的代价——老黄在时从没这样过。 4. 技术领先度没想象中大。 说领先5年?想多了。其实是不同领域领先几个月到一年不等,是老黄踩油门把节奏拉起来的。 更扎心的是,Anthropic最新路演显示:用AI agent把模型迁移到AMD GPU,只需3天,而以前要6-12个月。CUDA护城河正被AI coding agent动摇——老黄离开越久,衰退来得越快。 一句话总结:NVDA最大的护城河,就是老黄敢做高风险的架构决策,并能把资金、供应商、整个行业的信心拧到他那套愿景上。 现在连台积电、海力士都开始要求先付钱了——信心一旦松动,掉得比股价还快。 $NVDA @JensenHuang
黄仁勋离职开始反噬,台积电的做法很有意思,在黄总不当CEO之后,台积电出了个消息: 找中国台湾晶圆代工厂台积电排产能,以后可能先看账上有没有钱。业界传出,台积电 按各家公司的资产负债表拟一个信用额度。 超出信用额度的订单必须现金支付,很显然,一个会计当CEO,产业链是不太相信的。美国资本以为下了黄仁勋,就可以玩转整个AI产业链。 不到一周,在资本市场交出惨烈答卷之时,代工厂已经提示美国资本界,AI真正的掌门是谁。
62
178
10
1,282
261,781
Who Are Dario and Daniela Amodei? The Italian-American Siblings Behind Anthropic 🇮🇹🇺🇸 Both grew up in San Francisco’s Mission District Their father, Riccardo Amodei, was an Italian immigrant and leather craftsman, while their mother, Elena Engel, was an American of Jewish background Dario Amodei, born in 1983, followed an academic path in science After initially studying at Caltech, he graduated in physics from Stanford University and earned a PhD in biophysics at Princeton University as a Hertz Fellow He later completed postdoctoral research at Stanford School of Medicine He moved into artificial intelligence through research positions at Baidu and Google Brain, where he worked on deep learning and neural networks In 2016, he joined OpenAI, eventually becoming Vice President of Research He played a central role in the development of GPT-2 and GPT-3, contributed to research on scaling laws and co-invented Reinforcement Learning from Human Feedback (RLHF), a technique that uses human feedback to improve AI models Daniela Amodei, his younger sister, followed a different professional path Before entering the AI industry, she worked in public service and at Stripe, where she was Head of Risk Operations She later joined OpenAI, becoming Vice President of Safety and Policy, working on AI safety While Dario focused on scientific research and technical development, Daniela brought experience in operations, policy and company management In 2021, the two siblings left OpenAI alongside several other senior researchers and colleagues to co-found Anthropic Their decision followed disagreements over the direction of AI development and the importance given to long-term safety Anthropic was established as a Public Benefit Corporation, with a mission to develop AI systems that are reliable, interpretable and steerable Dario became CEO, leading research and the company’s technological direction, while Daniela became President and Chair of the Board, overseeing operations, strategy and business development Under their leadership, Anthropic developed the Claude family of AI models and grew into one of the industry’s leading companies, attracting major investments from some of the world’s largest technology companies Anthropic has just debuted on the stock market On November 19, 2026, the shares began trading on Nasdaq under the ticker ANT The IPO was priced at $83 per share (above the $70–$78 range), valuing the company at approximately $1.5 trillion on a fully diluted basis and raising $81.3 billion (plus a separate $26 billion private placement, for a total of about $107.3 billion) It is the largest IPO in history At the open, the stock debuted at $111.50 (roughly +34% above the IPO price) and is performing very strongly, with robust demand and a clearly positive first day of trading Dario and Daniela Amodei rang the opening bell at Nasdaq 🔔
175
318
47
2,147
421,302
un día como hoy fue escrita la mayor expresión de gratitud de la historia: la carta de Albert Camus a su profesor de la infancia
9
290
17
1,821
44,017
🏫Llega un nuevo colegio privado inglés a Aravaca para el proximo curso por 24000 €/año. Si en 2022/23 había dos colegios en ese rango de precios, este curso 2026/27 hay ya nueve que superan esa cifra. 🗨️urbanity.one/t/comunidad-de-…
4
10
72
14,796