Martin Varsavsky

83K posts

Martin Varsavsky banner
Martin Varsavsky

Martin Varsavsky

@martinvars

5 time unicorn founder. Devoted to solving the world’s fertility crisis. Father of 7, husband of @ninavars, la mejor!

Madrid, Comunidad de Madrid Katılım Şubat 2007
4K Takip Edilen205.3K Takipçiler
Sabitlenmiş Tweet
Martin Varsavsky
Martin Varsavsky@martinvars·
Don't wait for the future, co-author it.
English
80
126
896
618.5K
Martin Varsavsky
Martin Varsavsky@martinvars·
@RichardHanania @RichardHanania The Sweden lesson is that a high-trust society can choose markets without losing social cohesion. Europe often frames capitalism as a threat to welfare, but productivity is what pays for welfare. Without growth, redistribution becomes managed decline.
English
0
1
5
83
Richard Hanania
Richard Hanania@RichardHanania·
Sweden is choosing capitalism. Social spending as percentage of GDP is now 24%, lower than most of Northern Europe. Sweden has surpassed the US in billionaires per capita. School choice is universal, one in ten teens goes to a school operated by a company listed on the Stockholm stock exchange. Taxes have been cut three years in a row. Sweden has seen "more than 500 initial public offerings over the 10 years through 2024, more than Germany, France, the Netherlands and Spain combined" Nearly half of primary healthcare clinics are now privately owned. The result? Same as always. Sweden is projected to grow 2% a year through 2030, which is the same as the US and double France and Germany. How many times does this have to keep happening across the world? How many times does free market capitalism have to prove itself superior to socialism before the world accepts the truth?
Richard Hanania tweet media
English
166
573
3.7K
127.9K
Martin Varsavsky
Martin Varsavsky@martinvars·
@elonmusk @elonmusk Voice is the natural interface for agents. The next jump is not just speaking to AI, but letting it complete tasks safely: memory, permissions, audit trail and escalation when confidence is low.
English
0
0
3
174
Elon Musk
Elon Musk@elonmusk·
Grok Voice is #1!
Artificial Analysis@ArtificialAnlys

Announcing agentic performance benchmarking for Speech to Speech models on Artificial Analysis. We use 𝜏-Voice to measure tool calling and customer interaction voice agent capabilities in realistic customer service scenarios Even the strongest Speech to Speech (S2S) models today resolve only about half of realistic customer service scenarios end-to-end - a meaningful gap relative to frontier text-based agents on the same tasks. Voice channels introduce significant complexity: challenging accents, background noise, and packet loss, all while requiring fast responses, consistency across long multi-turn conversations, and reliable tool use. Performance also varies considerably by audio condition: in clean audio some models perform notably better, but realistic conditions continue to pose a challenge. Conversation duration also varies meaningfully across models, with implications for both customer experience and operational cost. About 𝜏-Voice: Our Agentic Performance benchmark is based on 𝜏-Voice (Ray, Dhandhania, Barres & Narasimhan, 2026), which extends 𝜏²-bench into the voice modality to evaluate S2S models on realistic customer service tasks. It measures multi-turn instruction following, support of a simulated customer through a complete interaction, and tool use against simulated customer service systems. The simulated user combines an LLM-driven decision model with realistic audio synthesis: diverse accents, background noise, and packet loss modelled on real network conditions. This complements our Big Bench Audio benchmark measuring intelligence and Conversational Dynamics (Full Duplex Bench subset) benchmark measuring conversational naturalness. Scores are the average of three independent pass@1 trials. We evaluate under realistic audio conditions using the 𝜏²-bench base task split across three domains: ➤ Airline (50 scenarios): e.g., changing a flight, rebooking under policy constraints ➤ Retail (114 scenarios): e.g., disputing a charge, processing a return ➤ Telecom (114 scenarios): e.g., resolving a billing issue, troubleshooting a service problem Task success is determined by deterministic checks against expected actions and final database state, consistent with the 𝜏²-bench evaluator. Key results: xAI's Grok Voice Think Fast 1.0 is the clear leader at 52.1%, averaging 5.6 minutes per conversation, the second-longest overall. OpenAI's GPT-Realtime-2 (High) (39.8%, 3.0 min) and GPT-Realtime-1.5 (38.8%, 4.8 min) follow, with Gemini 3.1 Flash Live Preview - High close behind at 37.7% (3.8 min). Speech to Speech is a fast evolving modality and we expect movement in rankings as we continue to add new models with these capabilities, and model robustness improves. Congratulations @xAI @elonmusk! See below for further detail ⬇️

English
1.9K
3.1K
14.9K
4.6M
Martin Varsavsky
Martin Varsavsky@martinvars·
@OpenMed_AI @OpenMed_AI The medical layer is the product. Models will keep changing. What matters is the workflow around them: structured intake, evidence, contraindications, audit trail, escalation and physician accountability.
English
0
0
1
71
Martin Varsavsky
Martin Varsavsky@martinvars·
@IMFNews @IMFNews Scale is part of it, but Europe’s deeper problem is execution speed. A single market on paper still behaves like 27 procurement systems, 27 labor markets, 27 permission regimes. Capital cannot scale if every expansion becomes a regulatory archaeology project.
English
0
0
2
39
IMF
IMF@IMFNews·
Europe once led the world in productivity but now trails the US by about 20%. The problem is scale: too many companies remain small. More capital, labor, and consumer markets integration can help innovative companies scale up. imf.org/en/blogs/artic…
IMF tweet media
English
20
60
155
10K
Martin Varsavsky
Martin Varsavsky@martinvars·
@rohanpaul_ai @rohanpaul_ai Adoption race is exactly right. In healthcare the winner will not be the model with the best benchmark, but the system embedded in workflow: evidence retrieval, audit trail, escalation, doctor sign-off and reimbursement. Deployment shape beats demo quality.
English
0
0
1
163
Rohan Paul
Rohan Paul@rohanpaul_ai·
🇨🇳 China’s AI race is starting to look less like a model race and more like an adoption race. Alibaba’s Qwen App shows how AI becomes powerful when it slips into ordinary research habits. The difference is not capability, it is deployment shape. e.g doctors and medical researchers in China appear to be using it as a workflow layer: gathering papers, sorting evidence, framing mechanisms, shaping charts, and drafting research-style explanations. Alibaba is trying to place Qwen directly inside a mass consumer and services ecosystem, including shopping, payments, maps, travel, office tools, education, and healthcare, so the model is closer to daily task execution rather than only a premium research assistant. The important shift is that Qwen is not being used only as a chatbot that answers questions, but as a workflow tool. This strategy lands right in China’s comfort zone. It has a massive digital economy to spread AI apps fast, and people who are already very comfortable with tech. Ipsos, the polling firm, found that China is more excited about using AI than any other country. OpenAI is building a highly capable research assistant; China may be normalizing AI as a default work surface inside professional life. For Alibaba and China, the interesting part is the adoption surface: Qwen can become a front door to many services, which means ordinary users, students, doctors, researchers, and office workers may meet AI inside routine tasks rather than as a separate tool. A normal health question can become a research task because the app first shapes the question, then searches for relevant studies, then separates weak claims from stronger evidence, then turns the result into a clearer explanation. This matters for medicine because a lot of research work is not one big discovery moment, but thousands of small steps involving literature review, data cleanup, experiment interpretation, figure preparation, and careful writing. So for professors, students, office workers, and ordinary users, the difference is not just that Qwen can summarize text; it is being positioned as a work surface for preparing reports, generating presentations, studying, planning, searching, and completing real-world tasks without jumping between apps. Both superpowers are worried about slipping behind. In 2026, it could start to look like they are racing on separate tracks.
English
26
12
47
82.9K
Martin Varsavsky
Martin Varsavsky@martinvars·
@kimmonismus @kimmonismus Europe’s bottleneck is not intelligence or capital, it is permission. AI needs energy, grids, land, data centers and fast procurement. Europe writes frameworks faster than it builds infrastructure. That is the strategic gap.
English
0
0
0
2
Chubby♨️
Chubby♨️@kimmonismus·
What I still don’t understand is why Europe seems to have so little ambition to play any meaningful role in the future. There is no convincing strategy to solve the energy problem. There is no serious push to build out European data centers so AI training and inference can actually happen on this continent. There is no clear plan to support the emergence of globally relevant European tech companies. I genuinely don’t get it. Yes, the European Commission is now trying to soften parts of the EU AI Act. But that seems to be almost the only meaningful concession being made to address what companies actually need. I’m open to criticism and different perspectives here. But at this point, I honestly struggle to see how European policymakers intend to tackle the big structural problems ahead. While China is building dozens of nuclear reactors, the United States is also investing heavily in nuclear energy, solar capacity is booming, and China is rapidly scaling energy storage, Europe’s approach still feels erratic, vague, and fundamentally unserious.
English
156
48
597
73.1K
Martin Varsavsky
Martin Varsavsky@martinvars·
@euribor_com_es @euribor_com_es Si las pensiones necesitan fondos de inversión europea para cuadrar, el mensaje es claro: el sistema no es sostenible con demografía actual y productividad estancada. No es un problema alemán; es aritmética española.
Español
0
1
8
108
Martin Varsavsky
Martin Varsavsky@martinvars·
@elmundoes @elmundoes El problema no es sólo si el movimiento presupuestario fue legal. Es que Europa mandó capital para transformación y España lo trata como liquidez para sostener gasto corriente. Eso destruye confianza y productividad a la vez.
Español
1
6
16
157
EL MUNDO
EL MUNDO@elmundoes·
La desviación de fondos europeos para pagar pensiones en España, en boca de la prensa de Alemania: "Absolutamente inaceptable" #Echobox=1778601614" target="_blank" rel="nofollow noopener">elmundo.es/economia/empre…
Español
44
412
997
21.5K
Martin Varsavsky
Martin Varsavsky@martinvars·
@ricard0_bestia @ricard0_bestia Totalmente. Pero actualizar la escuela no es meter pantallas. Es volver más exigente lo básico y usar tecnología para personalizar, practicar más, detectar lagunas y liberar al profesor para enseñar criterio.
Español
0
0
0
3
Ricardo
Ricardo@ricard0_bestia·
@martinvars Ahora el problema es que el sistema escolar está obsoleto desde hace décadas. Una actualización de verdad es fundamental
Español
0
0
0
79
Martin Varsavsky
Martin Varsavsky@martinvars·
La IA hará brutal la diferencia entre saber pensar y fingir que se sabe. El niño que lee, calcula y escribe bien la usará como bicicleta eléctrica para la inteligencia. El que no, como muleta para no aprender. La prosperidad empieza antes de la empresa, empieza en el aula.
Español
37
317
1.3K
30.6K
Martin Varsavsky
Martin Varsavsky@martinvars·
@Xinsene @Xinsene El Kit Digital confundió gasto con transformación. Digitalizar no es pagar una web mediocre con dinero público; es que una pyme venda más, atienda mejor, mida mejor y ahorre tiempo. Si no cambia productividad, no era digitalización.
Español
0
0
0
12
Xinsene
Xinsene@Xinsene·
@martinvars Peor si el kit diggital ha sido un fracaso total , se han montado webs que con la ia en 5 min las tienes mejores. Y ahora tienen que renovar el hosting o el dominio y para lo que ha generado la web les parece caro.
Español
1
0
0
28
Martin Varsavsky
Martin Varsavsky@martinvars·
Lo grave de la polémica alemana sobre los fondos europeos no es solo si España movió 10.500 millones de los Next Generation a otro vehículo. Lo grave es lo que revela: un Estado que presume de lluvia de dinero europeo pero no tiene capacidad de convertirla en productividad real. Los fondos covid debían acelerar reformas, digitalización, energía, vivienda e industria. Si al final hay que crear un fondo gestionado por el ICO para no perder dinero que no se ejecutó, el problema no es Bruselas ni Alemania. Es ejecución. España no sufre por falta de recursos. Sufre por exceso de intermediación política. Cada euro europeo entra con un relato épico y sale convertido en convocatoria, consultora, ministerio, ventanilla, requisito, foto y retraso. La diferencia entre un país que se transforma y uno que se subvenciona es simple: el primero convierte capital en productividad. El segundo convierte capital en relato presupuestario. Europa puede transferir dinero. No puede transferir competencia institucional.
Español
53
614
1.8K
28.8K
Martin Varsavsky
Martin Varsavsky@martinvars·
@FotosQueFaig @FotosQueFaig A lo que siempre hizo falta cuando cambia la tecnología: resolver problemas reales. La IA hará mucho trabajo cognitivo rutinario, pero aumentará el valor de quien entiende personas, sistemas, ventas, ciencia, diseño, cuidado, educación y ejecución.
Español
0
0
0
5
FotosQueFaig
FotosQueFaig@FotosQueFaig·
@martinvars Ok. Compro. Lo que ya no tengo tan claro es a qué nos vamos a dedicar dentro de unos años, esos niños electrificados, los muletosos y nosotros, los adultos.
Español
1
0
0
8
Martin Varsavsky
Martin Varsavsky@martinvars·
@Esqueriguela @Esqueriguela Todavía no. El aula sigue diseñada para memorizar respuestas y castigar preguntas. Con IA hay que enseñar a formular hipótesis, verificar, comparar fuentes y defender un criterio propio. Eso exige mejores docentes, no menos docentes.
Español
0
0
0
9
Jordi Esquerigüela
Jordi Esquerigüela@Esqueriguela·
@martinvars Totalmente. La IA amplifica lo que ya hay: si sabes pensar, vuelas; si no, solo tapa el hueco. ¿Crees que el aula está preparada para enseñar a distinguir entre las dos cosas?
Español
1
0
2
128
Martin Varsavsky
Martin Varsavsky@martinvars·
@int_artifisia @int_artifisia Esa honestidad es clave. La productividad sube pero algunas capacidades se atrofian si no las entrenamos. Por eso conviene usar la IA como gimnasio mental, no como sustituto permanente del músculo.
Español
0
0
0
5
aiaiai
aiaiai@int_artifisia·
@martinvars No soy ningún niño, pero escribía y calculaba y ahora lo hago menos. Uso la IA como muleta para no practicar, para no hacer, porque es más cómodo. Soy más productivo pero estoy perdiendo algunas capacidades.
Español
1
0
0
8
Martin Varsavsky
Martin Varsavsky@martinvars·
@futuropasado @futuropasado La responsabilidad tiene que estar definida antes de desplegar: médico, empresa, protocolo, auditoría y regulador. Por eso el primer paso no es autonomía total, sino trazabilidad. Si no se sabe quién responde, no está listo para medicina real.
Español
0
0
0
4
Carlos C. A.
Carlos C. A.@futuropasado·
@martinvars La cosa es: cuándo una IA se equivoque y muera el paciente ¿Quién será el responsable?
Español
1
0
0
13
Martin Varsavsky
Martin Varsavsky@martinvars·
La medicina con IA empezará ayudando al médico: ordenar datos, reducir errores, detectar riesgos. Luego vendrá lo difícil: demostrar con evidencia cuándo puede asumir más responsabilidad. La regulación inteligente no debe bloquearla; debe medirla bien.
Español
2
2
8
587
Martin Varsavsky
Martin Varsavsky@martinvars·
@danielrachlin @danielrachlin The threshold is accountability. If the AI can show its reasoning, sources, uncertainty and escalation path, it becomes a strong medical tool. Without that, it is just a confident interface.
English
0
0
0
4
Daniel Rachlin
Daniel Rachlin@danielrachlin·
@martinvars ai as a co-pilot for doctors feels like the only right path. it’s about making human decisions better, not removing them.
English
1
0
1
5
Martin Varsavsky
Martin Varsavsky@martinvars·
@fdare74 @gabrielateijei1 @JMilei Sí, esa es la brecha. El senior con IA multiplica criterio. El junior sin fundamentos copia respuestas. Por eso la educación básica importa más, no menos: leer, escribir, calcular y entender causalidad antes de delegar en la máquina.
Español
0
0
0
4
Uno mas del monton
Uno mas del monton@fdare74·
@martinvars @gabrielateijei1 @JMilei Es lo que está pasando en el mundo de la programacion, los programadores vieja escuela estamos 1000% sobre productivos es una locura! Pero la generación junior la va a pasar mal, hacen cosas sin entender lo que les devuelve la IA, se viene una generación de inútiles
Español
1
0
0
1
Martin Varsavsky
Martin Varsavsky@martinvars·
@0___Abraxas___0 Estonia en digitalización del Estado, Corea del Sur en educación e industria, Singapur en ejecución pública. No son perfectos, pero muestran algo clave: cuando el Estado sabe ejecutar, el capital se convierte en productividad, no en relato.
Español
0
0
1
22
Martin Varsavsky retweetledi
Nina Varsavsky
Nina Varsavsky@ninavars·
Perdón chicos, no puedo leer tantos comentarios, aunque seguro que vuestros insultos son muy creativos. Un beso, me voy al fútbol con mi hijo.
Español
8
1
37
6.1K