01 El miedo más antiguo del mundo The Oldest Fear in the World
Hay un miedo que acompaña a la humanidad desde que existe el progreso técnico. No es el miedo a la máquina en sí, sino algo más profundo: el miedo a volverse innecesario. La sensación de que un invento, un mecanismo, una forma nueva de hacer las cosas pueda dejar sin propósito a quien antes era imprescindible. Este miedo tiene siglos de antigüedad, y en cada época ha vestido ropajes distintos. Pero su esencia nunca cambia.
Para entenderlo, merece la pena recorrer las veces que la humanidad se enfrentó a él. Porque cada vez que ocurrió, la historia siguió un patrón tan consistente que resulta casi cómico que sigamos cayendo en la misma trampa.
Los destructores de máquinas
Inglaterra, 1811. En los condados del norte, donde la industria textil era el corazón económico de regiones enteras, corrían tiempos de angustia. Desde hacía décadas, máquinas cada vez más sofisticadas habían ido transformando los talleres artesanales. La Spinning Jenny de Hargreaves permitía hilar ocho hilos a la vez. La Water Frame de Arkwright hacía funcionar hiladoras con la fuerza del agua. Pero fue el telar mecánico de Cartwright el que cruzó un umbral psicológico: un solo operario con una de esas máquinas producía lo que antes requerían cuarenta tejedores manuales. Cuarenta familias cuyo sustento dependía de sus manos y su destreza, reemplazadas por un aparato de hierro y madera.
La reacción fue visceral. Grupos de trabajadores, autodenominados seguidores del mítico Ned Ludd, un personaje probablemente ficticio, comenzaron a destruir telares mecánicos de manera organizada. Enviaban cartas amenazantes a los dueños de fábricas. Atacaban de noche. El gobierno británico se lo tomó tan en serio que desplegó más soldados contra los luditas que los que el Duque de Wellington comandaba en la Península Ibérica luchando contra Napoleón. En 1812, el Parlamento aprobó la Frame Breaking Act, que establecía la pena de muerte por destruir maquinaria industrial.
La queja de los luditas era comprensible y humana: «Estas máquinas nos roban el sustento. Si las destruimos, preservamos nuestro modo de vida.» No eran ignorantes ni malvados. Eran personas que veían cómo su mundo se desmoronaba y hacían lo único que sabían hacer para defenderse.
Pero la historia les dio una respuesta que no esperaban. La ropa, que durante siglos había sido un bien semi-lujoso (la gente común poseía dos o tres prendas en toda su vida), se abarató tanto que surgió algo que antes no existía: la posibilidad de elegir qué ponerse. De tener ropa para diferentes ocasiones. De cambiar de guardarropa no por necesidad sino por gusto. De esa posibilidad nació la industria de la moda, el comercio textil a gran escala, las tiendas de ropa, la logística de distribución, la publicidad, la prensa especializada. En 1760 había unos 240.000 trabajadores textiles en Inglaterra. En 1830, a pesar de que cada máquina había reemplazado a docenas de artesanos, había 800.000. No trabajaban contra las máquinas: trabajaban con ellas, en una industria incomparablemente más grande que la que los luditas intentaron proteger.
Y el patrón no se detuvo ahí. Cada oleada tecnológica que vino después repitió la misma dinámica. Los caballos que movían ciudades enteras fueron reemplazados por el automóvil, y con ellos desaparecieron herreros, fabricantes de carruajes, establos y paleadores de estiércol, pero en su lugar emergió un universo de mecánicos, gasolineras, aseguradoras, constructoras de carreteras, moteles, turismo de masas y, según muchos historiadores, la propia clase media americana tal como la conocemos. La electricidad hizo desaparecer a los faroleros, a los repartidores de hielo, a los fabricantes de velas a gran escala, pero a cambio multiplicó las horas productivas del día, hizo posibles las fábricas de tres turnos, el entretenimiento nocturno, el cine, la radio, y transformó el hogar con electrodomésticos que liberaron millones de horas de trabajo doméstico hacia empleo remunerado y nuevas formas de consumo. En cada caso, las quejas fueron legítimas, el dolor de la transición fue real, y el resultado final fue una economía más grande, más diversa y más rica que la anterior.
En 1800, un trabajador inglés promedio dedicaba aproximadamente el ochenta por ciento de su ingreso a cubrir necesidades básicas: comida, ropa elemental, un techo. Para 1900 ese porcentaje había caído al cincuenta por ciento. Esa diferencia, un treinta por ciento del ingreso liberado, no desapareció. Se gastó en cosas que en 1800 ni siquiera tenían nombre: entretenimiento, educación, transporte, mobiliario, ocio. Industrias enteras nacieron del dinero que las máquinas liberaron al abaratar lo básico.
Cuando las máquinas empezaron a pensar
En 1952, el economista Wassily Leontief, futuro premio Nobel, formuló una analogía que aún hoy resuena con fuerza: «Las máquinas harán al trabajador lo que el tractor hizo al caballo.» La frase era potente y aterradora. El tractor efectivamente había hecho innecesarios a millones de caballos de trabajo. ¿Harían los ordenadores lo mismo con los humanos?
Los primeros ordenadores parecían darle la razón. Cuando IBM lanzó sus primeros sistemas comerciales en los años cincuenta y sesenta, los contables, mecanógrafos, operadores de centralita y calculistas humanos empezaron a verse amenazados. Un solo ordenador podía hacer en minutos lo que un equipo de veinte personas tardaba semanas. Las grandes empresas automatizaron sus departamentos de nóminas, de contabilidad, de gestión de inventarios. Los sindicatos protestaron. Los periódicos titularon sobre el fin del empleo administrativo. Los gobiernos crearon comisiones para estudiar el «problema de la automatización».
Y efectivamente, esos trabajos se destruyeron. El empleo industrial en Estados Unidos cayó del treinta y cinco al veinte por ciento del total entre 1960 y 1990. Las fábricas se automatizaron. Las oficinas se informatizaron. Mecanógrafas, archiveros, operadores de télex, dibujantes técnicos a mano: oficios que durante décadas habían sido la columna vertebral del empleo de clase media se evaporaron en una generación.
Pero lo que Leontief no pudo ver, como no lo pudieron ver los luditas ciento cuarenta años antes, fue lo que surgiría en su lugar. El sector servicios pasó de representar el sesenta por ciento del empleo al ochenta por ciento. Programadores, analistas de sistemas, diseñadores gráficos, gestores de bases de datos, técnicos de soporte informático: profesiones enteras que no existían antes de la revolución informática. Y no fueron solo trabajos técnicos. La informatización abarató los costes administrativos de crear y gestionar empresas, lo que desencadenó una oleada de emprendimiento. Pequeños negocios que antes no podían permitirse una contabilidad decente ahora podían gestionarse con un ordenador personal. La barrera de entrada a la actividad económica bajó, y con ella entraron millones de nuevos actores al mercado.
Y entonces llegó Internet. Y aceleró todo de una manera que nadie imaginó. Las agencias de viaje físicas se vaciaron. Los videoclubes cerraron: Blockbuster, con nueve mil tiendas en su mejor momento, desapareció en 2013. Kodak, que en 1998 empleaba a 145.000 personas, se declaró en bancarrota. Las tiendas de discos, los clasificados de periódico, las enciclopedias impresas, las páginas amarillas: todo arrasado en menos de una década. Las quejas fueron familiares y dolorosas: «Internet está destruyendo el comercio. Está acabando con industrias enteras.»
Pero mientras Blockbuster cerraba, Amazon crecía hasta emplear a más de un millón y medio de personas. Mientras Kodak quebraba, Instagram creaba una nueva economía visual. Mientras las enciclopedias morían, Wikipedia daba acceso gratuito al conocimiento humano y habilitaba una generación entera de autodidactas, investigadores y creadores. La economía de las aplicaciones, que no existía antes de 2008, generó millones de empleos directos e indirectos en todo el mundo. Los creadores de contenido, los especialistas en posicionamiento web, los científicos de datos, los gestores de comunidades: ninguna de estas profesiones tenía nombre en 1995. YouTube solo genera hoy más empleo que toda la industria del cine en 1990. La economía digital global se valoró en más de dieciséis billones de dólares en 2023, toda ella construida sobre las ruinas de industrias que Internet supuestamente destruyó.
Lo más notable del caso de los ordenadores e Internet es que la preocupación era exactamente la misma que con los telares: que la máquina haría innecesario al trabajador. Y la respuesta fue exactamente la misma: la máquina hizo innecesarios ciertos trabajos concretos, pero la riqueza que generó al abaratar bienes y servicios creó tal cantidad de actividad económica nueva que el empleo total en Estados Unidos pasó de setenta y ocho millones en 1970 a ciento sesenta millones en 2023. Un crecimiento de más del doble, mientras las máquinas supuestamente nos dejaban sin trabajo.
There is a fear that has accompanied humanity since technical progress began. It is not fear of the machine itself, but something deeper: the fear of becoming unnecessary. The feeling that an invention, a mechanism, a new way of doing things could render purposeless those who were once indispensable. This fear is centuries old, and in every era it has worn different clothes. But its essence never changes.
To understand it, it is worth retracing the occasions when humanity confronted it. Because each time it happened, history followed a pattern so consistent that it is almost comical that we keep falling into the same trap.
The Machine Breakers
England, 1811. In the northern counties, where the textile industry was the economic heart of entire regions, times of anguish prevailed. For decades, increasingly sophisticated machines had been transforming artisan workshops. Hargreaves's Spinning Jenny could spin eight threads at once. Arkwright's Water Frame powered spinning machines with the force of water. But it was Cartwright's power loom that crossed a psychological threshold: a single operator with one of those machines produced what had previously required forty manual weavers. Forty families whose livelihood depended on their hands and their skill, replaced by an apparatus of iron and wood.
The reaction was visceral. Groups of workers, self-proclaimed followers of the mythical Ned Ludd — a probably fictitious character — began destroying power looms in an organized fashion. They sent threatening letters to factory owners. They attacked at night. The British government took it so seriously that it deployed more soldiers against the Luddites than the Duke of Wellington commanded on the Iberian Peninsula fighting Napoleon. In 1812, Parliament passed the Frame Breaking Act, which established the death penalty for destroying industrial machinery.
The Luddites' complaint was understandable and human: "These machines are stealing our livelihood. If we destroy them, we preserve our way of life." They were neither ignorant nor evil. They were people who watched their world crumble and did the only thing they knew how to do to defend themselves.
But history gave them an answer they did not expect. Clothing, which for centuries had been a semi-luxury good (common people owned two or three garments in their entire lives), became so cheap that something emerged that had not existed before: the possibility of choosing what to wear. Of having clothes for different occasions. Of changing one's wardrobe not out of necessity but out of taste. From that possibility was born the fashion industry, large-scale textile trade, clothing stores, distribution logistics, advertising, specialized press. In 1760 there were about 240,000 textile workers in England. By 1830, despite each machine having replaced dozens of artisans, there were 800,000. They were not working against the machines: they were working with them, in an industry incomparably larger than the one the Luddites tried to protect.
And the pattern did not stop there. Every technological wave that followed repeated the same dynamic. The horses that moved entire cities were replaced by the automobile, and with them disappeared blacksmiths, carriage makers, stables, and manure shovelers, but in their place emerged a universe of mechanics, gas stations, insurance companies, road builders, motels, mass tourism, and, according to many historians, the American middle class itself as we know it. Electricity made lamplighters, ice deliverers, and large-scale candle manufacturers disappear, but in return it multiplied the productive hours of the day, made three-shift factories possible, enabled nighttime entertainment, cinema, radio, and transformed the home with appliances that freed millions of hours of domestic labor toward paid employment and new forms of consumption. In every case, the complaints were legitimate, the pain of the transition was real, and the final result was an economy larger, more diverse, and richer than the one before.
In 1800, an average English worker spent approximately eighty percent of their income covering basic necessities: food, basic clothing, a roof. By 1900, that percentage had fallen to fifty percent. That difference — thirty percent of income freed — did not disappear. It was spent on things that in 1800 did not even have names: entertainment, education, transportation, furniture, leisure. Entire industries were born from the money that machines freed by making the basics cheaper.
When Machines Began to Think
In 1952, the economist Wassily Leontief, future Nobel laureate, formulated an analogy that still resonates powerfully today: "Machines will do to the worker what the tractor did to the horse." The phrase was potent and terrifying. The tractor had indeed made millions of working horses unnecessary. Would computers do the same to humans?
The first computers seemed to prove him right. When IBM launched its first commercial systems in the fifties and sixties, accountants, typists, switchboard operators, and human calculators began to see themselves threatened. A single computer could do in minutes what a team of twenty people took weeks to accomplish. Large companies automated their payroll departments, accounting, inventory management. Unions protested. Newspapers ran headlines about the end of administrative employment. Governments created commissions to study the "automation problem."
And indeed, those jobs were destroyed. Industrial employment in the United States fell from thirty-five to twenty percent of the total between 1960 and 1990. Factories were automated. Offices were computerized. Typists, file clerks, telex operators, manual technical draftsmen: professions that for decades had been the backbone of middle-class employment evaporated in a single generation.
But what Leontief could not see, just as the Luddites could not see it a hundred and forty years before, was what would emerge in their place. The service sector went from representing sixty percent of employment to eighty percent. Programmers, systems analysts, graphic designers, database managers, IT support technicians: entire professions that did not exist before the computer revolution. And they were not only technical jobs. Computerization lowered the administrative costs of creating and managing businesses, which triggered a wave of entrepreneurship. Small businesses that previously could not afford decent accounting could now be managed with a personal computer. The barrier to entry into economic activity dropped, and with it millions of new players entered the market.
And then came the Internet. And it accelerated everything in a way nobody imagined. Physical travel agencies emptied out. Video rental stores closed: Blockbuster, with nine thousand stores at its peak, disappeared in 2013. Kodak, which in 1998 employed 145,000 people, filed for bankruptcy. Record stores, newspaper classifieds, printed encyclopedias, the Yellow Pages: all swept away in less than a decade. The complaints were familiar and painful: "The Internet is destroying commerce. It is wiping out entire industries."
But while Blockbuster was closing, Amazon was growing to employ more than one and a half million people. While Kodak was going bankrupt, Instagram was creating a new visual economy. While encyclopedias were dying, Wikipedia was providing free access to human knowledge and enabling an entire generation of autodidacts, researchers, and creators. The app economy, which did not exist before 2008, generated millions of direct and indirect jobs worldwide. Content creators, SEO specialists, data scientists, community managers: none of these professions had a name in 1995. YouTube alone now generates more employment than the entire film industry did in 1990. The global digital economy was valued at more than sixteen trillion dollars in 2023, all of it built on the ruins of industries that the Internet supposedly destroyed.
The most remarkable thing about the case of computers and the Internet is that the concern was exactly the same as with the looms: that the machine would make the worker unnecessary. And the answer was exactly the same: the machine made certain specific jobs unnecessary, but the wealth it generated by making goods and services cheaper created such an amount of new economic activity that total employment in the United States went from seventy-eight million in 1970 to one hundred and sixty million in 2023. Growth of more than double, while machines were supposedly leaving us without work.
Cada vez que la humanidad creyó que la máquina la condenaría, terminó más rica, más empleada y con más opciones que antes. El patrón no es coincidencia: es un principio económico. Every time humanity believed the machine would condemn it, it ended up richer, more employed, and with more options than before. The pattern is not a coincidence: it is an economic principle.Nous Capital Research · A Contra Corriente
Hay algo en esta repetición que debería hacernos reflexionar. No estamos ante una serie de casualidades afortunadas. Estamos ante un patrón tan consistente que exige una explicación. ¿Por qué las predicciones catastrofistas siempre se equivocan? ¿Qué mecanismo hace que la destrucción de empleos siempre genere más empleos nuevos? ¿Por qué la riqueza del mundo aumenta después de cada revolución tecnológica, en lugar de contraerse?
La respuesta no está en la tecnología. Está en la naturaleza de la economía y del ser humano. Y para entenderla, necesitamos dar un paso atrás y mirar los principios que gobiernan la actividad económica desde que existe.
There is something in this repetition that should make us reflect. We are not looking at a series of fortunate coincidences. We are looking at a pattern so consistent that it demands an explanation. Why do catastrophic predictions always prove wrong? What mechanism ensures that the destruction of jobs always generates more new ones? Why does global wealth increase after every technological revolution, instead of contracting?
The answer does not lie in technology. It lies in the nature of the economy and of the human being. And to understand it, we need to step back and look at the principles that have governed economic activity since it began.
02 La falacia del pastel fijo The Fixed Pie Fallacy
Antes de sumergirnos en los principios económicos, conviene detenernos en el error que conecta a los luditas de 1811 con los titulares alarmistas de hoy sobre la inteligencia artificial. Es un error tan antiguo como persistente, y tiene un nombre: la falacia del pastel fijo.
El razonamiento funciona así: la economía es como un pastel de tamaño determinado. Hay un número fijo de empleos, una cantidad fija de riqueza, un volumen fijo de trabajo por hacer. Si una máquina se lleva una porción, a ti te queda menos. Si el telar mecánico hace el trabajo de cuarenta tejedores, treinta y nueve se quedan sin porción. Si la inteligencia artificial redacta contratos, los abogados junior pierden la suya. Parece lógico. Parece evidente. Y es profundamente falso.
El economista francés Frédéric Bastiat lo entendió mejor que nadie. En 1850 publicó un ensayo breve y devastador titulado Lo que se ve y lo que no se ve, donde expuso una idea que debería ser lectura obligatoria para cualquiera que opine sobre tecnología y empleo:
Before diving into economic principles, it is worth pausing at the error that connects the Luddites of 1811 with today's alarmist headlines about artificial intelligence. It is an error as old as it is persistent, and it has a name: the fixed pie fallacy.
The reasoning goes like this: the economy is like a pie of a fixed size. There is a fixed number of jobs, a fixed amount of wealth, a fixed volume of work to be done. If a machine takes a slice, you are left with less. If the power loom does the work of forty weavers, thirty-nine are left without a slice. If artificial intelligence drafts contracts, junior lawyers lose theirs. It seems logical. It seems obvious. And it is profoundly false.
The French economist Frederic Bastiat understood this better than anyone. In 1850, he published a brief and devastating essay titled That Which Is Seen, and That Which Is Not Seen, in which he laid out an idea that should be required reading for anyone who offers opinions on technology and employment:
"The bad economist confines himself to the visible effect; the good economist takes into account both the effect that can be seen and those effects that must be foreseen. The bad economist pursues a small present good that will be followed by a great evil to come. The good economist pursues a great good to come, at the risk of a small present evil."
Frederic Bastiat, That Which Is Seen, and That Which Is Not Seen, 1850
La genialidad de Bastiat fue señalar que en toda transformación económica hay efectos visibles y efectos invisibles. Los visibles son inmediatos, concretos, fáciles de fotografiar. Los invisibles son futuros, difusos, imposibles de imaginar hasta que ocurren. Y el problema es que los seres humanos estamos programados para prestar atención a lo visible y descartar lo invisible.
Cuando el telar mecánico reemplazó a los tejedores, lo visible era desgarrador: familias sin sustento, comunidades rotas, un oficio centenario extinguido. Lo invisible era la industria de la moda, el comercio textil global, los millones de empleos en retail y logística que surgirían de la ropa barata. Cuando Internet destruyó los videoclubes, lo visible fue Blockbuster cerrando nueve mil tiendas. Lo invisible fue Netflix, YouTube, Spotify, el ecosistema de creadores de contenido, la economía del streaming que hoy vale más que todos los videoclubes que existieron jamás.
Hoy estamos en un momento similar, quizá más amplio que cualquiera de los anteriores. La inteligencia artificial no se limita a abaratar un sector específico. Está abaratando capas enteras de la estructura económica que hasta ahora funcionaban como barreras de entrada y costes fijos para cualquier actividad productiva. Los sistemas de información, que requerían departamentos enteros de analistas, están siendo comprimidos a una fracción de su coste. Los trámites legales y regulatorios, que exigían abogados, consultores y meses de gestión, empiezan a resolverse en horas con asistentes inteligentes. El análisis de inversiones, que requería equipos de expertos financieros, es cada vez más accesible a través de herramientas automatizadas. La barrera regulatoria misma, esa maraña de requisitos que impedía a las pequeñas empresas competir con las grandes, se está erosionando porque la IA permite cumplir con ella a una fracción del coste anterior. El software, el análisis de datos, la enseñanza, la traducción, la contabilidad: muchas de las capas de trabajo que sostienen nuestra economía se están abaratando simultáneamente.
Esto genera angustia, y es natural. La parte visible de la ecuación es real: trabajos concretos se transformarán, habilidades concretas perderán su valor de mercado, transiciones concretas serán dolorosas. Pero la parte invisible también será real, aunque hoy no podamos fotografiarla. Si todo eso se abarata, el dinero que la sociedad se ahorra no desaparece del sistema. Se redirige. Fluye hacia necesidades que hoy ni siquiera tienen nombre, como no tenía nombre en 1995 la profesión de «creador de contenido» ni en 2005 la de «ingeniero de prompts».
El pastel no tiene un tamaño fijo. Cada revolución tecnológica ha demostrado que el pastel crece. Y crece precisamente porque la tecnología libera recursos que se redirigen hacia actividades nuevas que antes eran imposibles o impensables. La pregunta correcta, la que Bastiat nos enseñó a hacer, no es si la IA destruirá empleos. Los destruirá. La pregunta correcta es: ¿qué se creará con los recursos que la IA libere? Y si la historia sirve de guía, la respuesta es: más de lo que hoy podemos imaginar.
Bastiat's genius was in pointing out that in every economic transformation there are visible effects and invisible effects. The visible ones are immediate, concrete, easy to photograph. The invisible ones are future, diffuse, impossible to imagine until they happen. And the problem is that human beings are wired to pay attention to the visible and dismiss the invisible.
When the power loom replaced weavers, what was visible was heartbreaking: families without a livelihood, broken communities, a centuries-old trade extinguished. What was invisible was the fashion industry, global textile commerce, the millions of jobs in retail and logistics that would emerge from cheap clothing. When the Internet destroyed video rental stores, what was visible was Blockbuster closing nine thousand stores. What was invisible was Netflix, YouTube, Spotify, the content creator ecosystem, the streaming economy that today is worth more than all the video rental stores that ever existed.
Today we are at a similar moment, perhaps broader than any before. Artificial intelligence is not limited to making a specific sector cheaper. It is making entire layers of the economic structure cheaper — layers that until now functioned as barriers to entry and fixed costs for any productive activity. Information systems, which required entire departments of analysts, are being compressed to a fraction of their cost. Legal and regulatory procedures, which demanded lawyers, consultants, and months of management, are beginning to be resolved in hours with intelligent assistants. Investment analysis, which required teams of financial experts, is becoming increasingly accessible through automated tools. The regulatory barrier itself, that tangle of requirements that prevented small companies from competing with large ones, is being eroded because AI allows compliance at a fraction of the previous cost. Software, data analysis, teaching, translation, accounting: many of the layers of work that sustain our economy are becoming cheaper simultaneously.
This generates anxiety, and that is natural. The visible part of the equation is real: specific jobs will be transformed, specific skills will lose their market value, specific transitions will be painful. But the invisible part will also be real, even if we cannot photograph it today. If all of that becomes cheaper, the money that society saves does not disappear from the system. It gets redirected. It flows toward needs that today do not even have names, just as the profession of "content creator" had no name in 1995, nor "prompt engineer" in 2005.
The pie does not have a fixed size. Every technological revolution has demonstrated that the pie grows. And it grows precisely because technology frees resources that are redirected toward new activities that were previously impossible or unthinkable. The right question, the one Bastiat taught us to ask, is not whether AI will destroy jobs. It will. The right question is: what will be created with the resources that AI frees? And if history is any guide, the answer is: more than we can imagine today.
03 Los principios que gobiernan la economía The Principles That Govern the Economy
La historia que acabamos de recorrer muestra un patrón claro, pero no lo explica. Para entender por qué cada revolución tecnológica ha acabado generando más riqueza y más empleo que los que destruyó, necesitamos ir a los principios fundamentales que gobiernan la actividad económica. No son principios inventados por académicos en una torre de marfil: son observaciones sobre la naturaleza humana y la escasez que se cumplen con la regularidad de una ley física. La Escuela Austriaca de Economía, fundada por Carl Menger en 1871 y desarrollada por pensadores como Ludwig von Mises y Friedrich Hayek, construyó el marco teórico más sólido para comprenderlos.
Todo empieza por la acción humana
El punto de partida de todo el edificio es una observación tan simple que parece banal, hasta que se entienden sus consecuencias. Ludwig von Mises, en su obra magna La Acción Humana (1949), la formuló así:
The history we have just surveyed shows a clear pattern, but it does not explain it. To understand why every technological revolution has ended up generating more wealth and more employment than it destroyed, we need to go to the fundamental principles that govern economic activity. These are not principles invented by academics in an ivory tower: they are observations about human nature and scarcity that hold with the regularity of a physical law. The Austrian School of Economics, founded by Carl Menger in 1871 and developed by thinkers such as Ludwig von Mises and Friedrich Hayek, built the most robust theoretical framework for understanding them.
It All Begins with Human Action
The starting point of the entire edifice is an observation so simple it seems banal — until one understands its consequences. Ludwig von Mises, in his magnum opus Human Action (1949), formulated it as follows:
"Human action is purposeful behavior. Or we may say: Action is will put into operation and transformed into an agency, is aiming at ends and goals."
Ludwig von Mises, Human Action, 1949
Parece obvio. Pero detrás de esta definición hay una lógica implacable. Para que un ser humano actúe, para que haga cualquier cosa, se necesitan exactamente tres condiciones.
Primero, insatisfacción: el individuo no está plenamente contento con su situación actual. Siente que algo falta, que algo podría ser mejor, que hay un hueco entre lo que tiene y lo que desearía tener.
Segundo, la imagen de un estado mejor: puede imaginar una situación más satisfactoria que la presente. No basta con estar insatisfecho; necesita concebir una alternativa.
Tercero, la capacidad percibida de actuar: cree que puede hacer algo para pasar del estado actual al estado deseado. No basta con imaginar algo mejor; necesita sentir que tiene medios para alcanzarlo.
Pensemos ahora qué tendría que ocurrir para que la actividad económica cesara. Tendría que desaparecer al menos una de estas tres condiciones. Que el ser humano estuviera plenamente satisfecho, sin deseo alguno de mejora. O que no pudiera imaginar nada mejor que su situación presente. O que se sintiera completamente impotente para cambiar nada.
La primera posibilidad no se ha dado jamás en la historia de la humanidad. Vivimos materialmente mejor que cualquier generación anterior, y sin embargo nuestro nivel de insatisfacción, nuestro motor económico, no ha disminuido un ápice. Hoy nos preocupa la velocidad del wifi, la calidad del café, si nuestro trabajo es suficientemente «significativo», si nuestra alimentación es orgánica, si nuestra huella de carbono es aceptable. Preocupaciones que habrían sido incomprensibles para alguien cuya prioridad era sobrevivir al próximo invierno. Cada necesidad satisfecha abre la puerta a una nueva más sofisticada. Es una escalera que no tiene último peldaño.
La segunda posibilidad es refutada por la historia entera de la civilización. Cada generación ha imaginado mundos que la anterior no podía concebir.
La tercera posibilidad no solo no se cumple con la inteligencia artificial, sino que ocurre exactamente lo contrario: la IA amplifica nuestra capacidad de actuar. Nos da herramientas más poderosas que nunca para transformar nuestra insatisfacción en acción.
It seems obvious. But behind this definition lies an implacable logic. For a human being to act, to do anything at all, exactly three conditions are required.
First, dissatisfaction: the individual is not fully content with their current situation. They feel that something is missing, that something could be better, that there is a gap between what they have and what they would like to have.
Second, the image of a better state: they can imagine a situation more satisfactory than the present one. It is not enough to be dissatisfied; they need to conceive of an alternative.
Third, the perceived capacity to act: they believe they can do something to move from the current state to the desired one. It is not enough to imagine something better; they need to feel they have the means to achieve it.
Now let us consider what would have to happen for economic activity to cease. At least one of these three conditions would have to disappear. That the human being be fully satisfied, with no desire for improvement. Or that they could imagine nothing better than their present situation. Or that they felt completely powerless to change anything.
The first possibility has never occurred in the history of humanity. We live materially better than any previous generation, and yet our level of dissatisfaction — our economic engine — has not diminished one iota. Today we worry about wifi speed, the quality of our coffee, whether our work is sufficiently "meaningful," whether our diet is organic, whether our carbon footprint is acceptable. Concerns that would have been incomprehensible to someone whose priority was surviving the next winter. Every need satisfied opens the door to a more sophisticated new one. It is a staircase that has no final step.
The second possibility is refuted by the entire history of civilization. Every generation has imagined worlds that the previous one could not conceive.
The third possibility is not only unfulfilled by artificial intelligence, but exactly the opposite occurs: AI amplifies our capacity to act. It gives us more powerful tools than ever to transform our dissatisfaction into action.
Aplicado directamente a la inteligencia artificial: la IA no reduce la acción humana, la potencia. El programador que antes necesitaba semanas para construir una aplicación puede hacerlo en horas. El investigador que tardaba meses en analizar datos puede hacerlo en minutos. El emprendedor que no podía costear un equipo legal ahora accede a asesoramiento inteligente. En cada caso, la IA no elimina la acción sino que amplía el rango de lo que cada persona puede intentar. Más capacidad de actuar significa más acción, y más acción significa más economía, no menos.
El valor está en los ojos de quien mira
Si la acción humana es el motor, necesitamos entender qué la dirige. ¿Cómo decide el ser humano hacia dónde actuar? ¿Qué determina que un bien valga más que otro?
Carl Menger, el padre de la Escuela Austriaca, dio la respuesta en 1871 con su teoría del valor subjetivo. Antes de Menger, la idea dominante, defendida por pensadores tan influyentes como Karl Marx, era que el valor de un bien dependía del trabajo invertido en producirlo. Cuantas más horas de esfuerzo, más valor. Menger demostró que esto era un error fundamental:
Applied directly to artificial intelligence: AI does not reduce human action; it empowers it. The programmer who previously needed weeks to build an application can do it in hours. The researcher who took months to analyze data can do it in minutes. The entrepreneur who could not afford a legal team now accesses intelligent counsel. In every case, AI does not eliminate action but expands the range of what each person can attempt. More capacity to act means more action, and more action means more economy, not less.
Value Is in the Eye of the Beholder
If human action is the engine, we need to understand what directs it. How does a human being decide where to act? What determines that one good is worth more than another?
Carl Menger, the father of the Austrian School, gave the answer in 1871 with his theory of subjective value. Before Menger, the dominant idea, defended by thinkers as influential as Karl Marx, was that the value of a good depended on the labor invested in producing it. The more hours of effort, the more value. Menger demonstrated that this was a fundamental error:
"The value of goods arises from their relationship to our needs, and is not inherent in the goods themselves. With changes in this relationship, value also changes."
Carl Menger, Principles of Economics, 1871
El valor no está en el objeto: está en la mente de quien lo necesita. Un vaso de agua no tiene un valor intrínseco; tiene el valor que tú le das según tu sed. En un río caudaloso, el agua carece prácticamente de valor. En un desierto, darías todo lo que posees por un litro. El agua es la misma; lo que cambia es la relación entre tu necesidad y la cantidad disponible.
Esta idea tiene una consecuencia enorme: es el consumidor quien decide cómo se reestructura el mercado y la economía. No lo decide el productor, ni el empresario, ni la tecnología. Lo decide la persona que compra, que elige, que valora. Cada acto de compra es un voto sobre qué merece existir y qué no. Y esa votación agregada de millones de consumidores es la fuerza que moldea la economía, por encima de cualquier otra.
Pensemos en lo que esto implica para la inteligencia artificial. El valor de la IA como factor productivo no lo determina su capacidad técnica. Lo determina el consumidor. Si las personas deciden que valoran más un texto escrito por un ser humano que uno generado por IA, entonces el texto humano tendrá más valor en el mercado, independientemente de que la IA sea técnicamente capaz de producirlo más rápido y más barato. Si los consumidores empiezan a valorar la autenticidad, la conexión personal, el «hecho por humanos» como una cualidad deseable, entonces la IA pierde valor como factor productivo en esos sectores, por mucho que sea técnicamente superior. Ya lo vemos hoy en fenómenos como la comida artesanal, la cerveza artesana o la música en directo: el consumidor paga más por algo «menos eficiente» porque valora precisamente lo que la máquina no puede aportar. El valor no está en la eficiencia de la producción sino en la satisfacción que el consumidor obtiene, y esa satisfacción la define él, no la máquina.
La escasez: el concepto que lo explica todo
Del valor subjetivo se deriva el concepto más poderoso de toda la economía: la escasez. Menger distinguió con una claridad asombrosa entre dos tipos de bienes.
Los bienes libres son aquellos cuya cantidad disponible excede la necesidad que los seres humanos tienen de ellos. El aire que respiramos, la luz del sol, la arena en un desierto. Nadie pelea por ellos, nadie los intercambia, nadie paga por ellos. Simplemente hay suficiente para todos.
Los bienes económicos son aquellos cuya cantidad disponible es inferior a la necesidad. La tierra fértil, los minerales, el tiempo de un cirujano. Hay más gente que los quiere que cantidad disponible. Y de esa brecha entre deseo y disponibilidad nace todo: el precio, el intercambio, la competencia, la propiedad, el mercado. En una palabra: la economía.
Pero la observación verdaderamente revolucionaria de Menger fue que un mismo bien puede migrar entre ambas categorías. No es una propiedad fija del bien; es una relación entre la oferta y la necesidad humana que cambia con las circunstancias. El agua es un bien libre junto a un manantial y un bien económico en el desierto. La información era un bien económico escaso cuando había que comprar enciclopedias; se convirtió en cuasi-libre con Wikipedia y Google. La privacidad era un bien libre cuando el anonimato era lo natural; se convirtió en un bien económico multimillonario cuando Internet la puso en peligro.
Y aquí está la clave de todo nuestro argumento: cada vez que la tecnología convierte un bien económico en un bien cuasi-libre, dos cosas ocurren simultáneamente. La primera es visible: el bien se abarata y quienes vivían de producirlo pierden su ventaja. La segunda es invisible pero más importante: emergen nuevas escaseces. Necesidades que antes no existían o no se percibían se vuelven relevantes, y al ser escasas, se convierten en nuevos bienes económicos. Cuando la información se hizo abundante, la curación de información se volvió escasa. Cuando la comunicación se hizo gratuita, la atención de las personas se convirtió en el recurso más cotizado del planeta. La escasez no desaparece con la tecnología. Se desplaza, se transforma, migra de unos bienes a otros. Pero la brecha entre lo que los humanos desean y lo que está disponible no se cierra jamás, porque los deseos humanos crecen al menos tan rápido como la capacidad de satisfacerlos.
La IA está acelerando esta dinámica a una escala sin precedentes. Está convirtiendo en abundantes bienes que hasta ayer eran costosos: la traducción, el código básico, el contenido digital, el asesoramiento genérico. Eso es lo que se ve. Lo que no se ve todavía, pero la lógica económica garantiza, es la emergencia de nuevas escaseces: la autenticidad verificada en un mundo de contenido generado, la atención humana genuina en un mundo de interacciones artificiales, la curación experta en un mundo de producción infinita, la confianza verificable en un mundo de deepfakes. Cada uno de estos será un nuevo bien económico, con su propio mercado, sus propios precios, sus propios empleos. La economía no se vacía: su centro de gravedad se desplaza.
Donde hay escasez, hay propiedad
De esta observación sobre la escasez se deriva otra de enorme importancia. Menger argumentó que la propiedad privada no es una invención legal arbitraria, ni una imposición del poder, ni un acuerdo social que podríamos revocar si quisiéramos. Es la consecuencia lógica e inevitable de la escasez:
Value is not in the object: it is in the mind of the person who needs it. A glass of water has no intrinsic value; it has the value you assign it based on your thirst. By a rushing river, water has practically no value. In a desert, you would give everything you own for a liter. The water is the same; what changes is the relationship between your need and the available quantity.
This idea has an enormous consequence: it is the consumer who decides how the market and the economy are restructured. Neither the producer, nor the entrepreneur, nor the technology decides. The person who buys, who chooses, who values, decides. Every act of purchase is a vote on what deserves to exist and what does not. And that aggregated vote of millions of consumers is the force that shapes the economy, above all else.
Let us think about what this implies for artificial intelligence. The value of AI as a factor of production is not determined by its technical capability. It is determined by the consumer. If people decide they value a text written by a human more than one generated by AI, then the human text will have more value in the market, regardless of the fact that AI is technically capable of producing it faster and cheaper. If consumers begin to value authenticity, personal connection, and "made by humans" as a desirable quality, then AI loses value as a factor of production in those sectors, no matter how technically superior it may be. We already see this today in phenomena such as artisanal food, craft beer, or live music: the consumer pays more for something "less efficient" because they value precisely what the machine cannot provide. Value does not reside in the efficiency of production but in the satisfaction the consumer obtains, and that satisfaction is defined by the consumer, not by the machine.
Scarcity: The Concept That Explains Everything
From subjective value derives the most powerful concept in all of economics: scarcity. Menger distinguished with astonishing clarity between two types of goods.
Free goods are those whose available quantity exceeds the need that human beings have for them. The air we breathe, sunlight, sand in a desert. No one fights over them, no one trades them, no one pays for them. There is simply enough for everyone.
Economic goods are those whose available quantity is less than the need. Fertile land, minerals, a surgeon's time. There are more people who want them than there is available supply. And from that gap between desire and availability everything is born: price, exchange, competition, property, the market. In a word: the economy.
But Menger's truly revolutionary observation was that one and the same good can migrate between both categories. It is not a fixed property of the good; it is a relationship between supply and human need that changes with circumstances. Water is a free good by a spring and an economic good in the desert. Information was a scarce economic good when you had to buy encyclopedias; it became quasi-free with Wikipedia and Google. Privacy was a free good when anonymity was the natural state; it became a multimillion-dollar economic good when the Internet put it in jeopardy.
And here lies the key to our entire argument: every time technology converts an economic good into a quasi-free good, two things happen simultaneously. The first is visible: the good becomes cheaper and those who made their living producing it lose their advantage. The second is invisible but more important: new scarcities emerge. Needs that did not previously exist or were not perceived become relevant, and being scarce, they become new economic goods. When information became abundant, the curation of information became scarce. When communication became free, people's attention became the most valued resource on the planet. Scarcity does not disappear with technology. It shifts, transforms, migrates from some goods to others. But the gap between what humans desire and what is available never closes, because human desires grow at least as fast as the capacity to satisfy them.
AI is accelerating this dynamic on an unprecedented scale. It is making abundant goods that until yesterday were costly: translation, basic code, digital content, generic advice. That is what is visible. What is not yet visible, but economic logic guarantees, is the emergence of new scarcities: verified authenticity in a world of generated content, genuine human attention in a world of artificial interactions, expert curation in a world of infinite production, verifiable trust in a world of deepfakes. Each of these will be a new economic good, with its own market, its own prices, its own jobs. The economy does not empty out: its center of gravity shifts.
Where There Is Scarcity, There Is Property
From this observation about scarcity derives another of enormous importance. Menger argued that private property is not an arbitrary legal invention, nor an imposition of power, nor a social agreement that we could revoke if we wished. It is the logical and inevitable consequence of scarcity:
"Where the quantity available of a good is insufficient to satisfy all needs, competition for its possession necessarily arises, and from this competition property emerges."
Carl Menger, Principles of Economics, 1871
Si un recurso es abundante para todos, no hace falta asignarlo. No hay propiedad sobre el aire que respiramos porque hay suficiente para todos. Pero en el momento en que la demanda supera la oferta, alguien tiene que decidir quién lo obtiene, cuánto y a cambio de qué. Y de esa necesidad de asignación nace la propiedad.
La consecuencia es directa: mientras la tecnología siga generando nuevas escaseces, y la historia demuestra que siempre lo hace, habrá propiedad. Y donde hay propiedad hay derechos que defender, intercambios que realizar, precios que negociar, contratos que cumplir. Hay actividad económica.
Podemos completar ahora la cadena lógica. Los deseos humanos insatisfechos generan acción (Mises). Esa acción se dirige hacia bienes que el individuo valora subjetivamente (Menger). Cuando esos bienes son escasos, surgen la propiedad y el intercambio. Y como la IA genera nuevas escaseces al mismo tiempo que resuelve las antiguas, la cadena no se rompe: se reconfigura. El código básico deja de ser escaso, pero la arquitectura de software confiable sigue siéndolo. La información genérica deja de ser escasa, pero el criterio para filtrar lo relevante sigue siéndolo. Y donde algo sigue siendo escaso, alguien lo poseerá, alguien lo intercambiará, alguien pagará por ello.
El problema del conocimiento
Friedrich Hayek, discípulo de Mises y premio Nobel de Economía en 1974, añadió una pieza fundamental al edificio. En 1945 publicó un artículo breve y extraordinariamente denso titulado El uso del conocimiento en la sociedad, donde formuló una idea que es quizá la más relevante de toda la ciencia económica para entender los límites de cualquier inteligencia, artificial o humana:
If a resource is abundant for everyone, there is no need to allocate it. There is no property over the air we breathe because there is enough for everyone. But the moment demand exceeds supply, someone has to decide who gets it, how much, and in exchange for what. And from that need for allocation, property is born.
The consequence is direct: as long as technology continues to generate new scarcities — and history demonstrates that it always does — there will be property. And where there is property, there are rights to defend, exchanges to make, prices to negotiate, contracts to fulfill. There is economic activity.
We can now complete the logical chain. Unsatisfied human desires generate action (Mises). That action is directed toward goods that the individual values subjectively (Menger). When those goods are scarce, property and exchange arise. And since AI generates new scarcities at the same time it resolves old ones, the chain does not break: it reconfigures. Basic code ceases to be scarce, but reliable software architecture remains so. Generic information ceases to be scarce, but the judgment to filter what is relevant remains so. And where something remains scarce, someone will own it, someone will trade it, someone will pay for it.
The Knowledge Problem
Friedrich Hayek, a disciple of Mises and Nobel laureate in Economics in 1974, added a fundamental piece to the edifice. In 1945, he published a brief and extraordinarily dense article titled The Use of Knowledge in Society, in which he formulated an idea that is perhaps the most relevant in all of economic science for understanding the limits of any intelligence, artificial or human:
"The economic problem of society is not merely a problem of how to allocate 'given' resources. It is a problem of how to secure the best use of resources known to any of the members of society, for ends whose relative importance only these individuals know."
Friedrich Hayek, The Use of Knowledge in Society, 1945
Lo que Hayek señaló es que el conocimiento relevante para coordinar una economía no está en un lugar centralizado. Está disperso entre millones de personas, y gran parte de él es de una naturaleza que lo hace imposible de centralizar. El comerciante local que sabe que su barrio prefiere cierto tipo de producto. El ingeniero de planta que percibe que una máquina suena distinto y va a fallar. El consumidor que esta mañana decidió que prefiere arroz a pasta, sin poder explicar exactamente por qué. Todo este conocimiento es tácito, no se puede verbalizar completamente; es contextual, depende de circunstancias específicas de tiempo y lugar; y es dinámico, cambia constantemente.
¿Cómo se coordina una sociedad entera cuando el conocimiento necesario para hacerlo está fragmentado en millones de mentes? La respuesta de Hayek fue: mediante el sistema de precios. Cuando un bien se vuelve más escaso, su precio sube, y esa subida transmite una señal a millones de personas simultáneamente: economiza este recurso, busca alternativas, invierte en producir más. Nadie necesita saber por qué subió el precio. El precio mismo contiene la información necesaria. Es un sistema de coordinación descentralizado, involuntario y extraordinariamente eficiente que ningún planificador central ha podido replicar.
Para la inteligencia artificial, esto tiene una implicación profunda. Por mucha capacidad de cálculo que tenga una IA, el problema económico no es un problema de cálculo: es un problema de conocimiento. Los precios de mercado no son números que un planificador pueda replicar; son señales que condensan las valoraciones subjetivas de millones de personas en cada instante. Una IA puede calcular a velocidades inimaginables, pero no puede saber qué valora cada persona en cada momento sin un sistema de precios que lo revele. El mercado no es un sistema imperfecto que la tecnología reemplazará: es el mecanismo que genera la información misma que cualquier sistema de coordinación necesita.
El consumidor dirige la economía
Hay un principio que se desprende de todo lo anterior y que es crucial para entender hacia dónde se dirigirá la economía en la era de la inteligencia artificial: la soberanía del consumidor.
Mises insistió a lo largo de toda su obra en que, en una economía de mercado, no son los productores ni los empresarios quienes dirigen la economía. Son los consumidores. Cada acto de compra es un voto. Cada decisión de gastar el dinero en una cosa y no en otra envía una señal al sistema productivo sobre qué se valora y qué no. Los empresarios que interpretan correctamente estas señales prosperan. Los que no, quiebran.
What Hayek pointed out is that the knowledge relevant to coordinating an economy is not located in a centralized place. It is dispersed among millions of people, and much of it is of a nature that makes it impossible to centralize. The local shopkeeper who knows that their neighborhood prefers a certain type of product. The plant engineer who perceives that a machine sounds different and is about to fail. The consumer who this morning decided they prefer rice to pasta, without being able to explain exactly why. All this knowledge is tacit — it cannot be fully verbalized; it is contextual — it depends on specific circumstances of time and place; and it is dynamic — it changes constantly.
How does an entire society coordinate when the knowledge needed to do so is fragmented across millions of minds? Hayek's answer was: through the price system. When a good becomes scarcer, its price rises, and that rise transmits a signal to millions of people simultaneously: economize on this resource, look for alternatives, invest in producing more. Nobody needs to know why the price rose. The price itself contains the necessary information. It is a decentralized, involuntary, and extraordinarily efficient coordination system that no central planner has been able to replicate.
For artificial intelligence, this has a profound implication. No matter how much computing power an AI has, the economic problem is not a problem of computation: it is a problem of knowledge. Market prices are not numbers that a planner can replicate; they are signals that condense the subjective valuations of millions of people at every instant. An AI can calculate at unimaginable speeds, but it cannot know what each person values at each moment without a price system to reveal it. The market is not an imperfect system that technology will replace: it is the mechanism that generates the very information that any coordination system needs.
The Consumer Directs the Economy
There is a principle that follows from all of the above and that is crucial for understanding where the economy will head in the era of artificial intelligence: consumer sovereignty.
Mises insisted throughout his entire body of work that, in a market economy, it is neither the producers nor the entrepreneurs who direct the economy. It is the consumers. Every act of purchase is a vote. Every decision to spend money on one thing and not another sends a signal to the productive system about what is valued and what is not. The entrepreneurs who correctly interpret these signals prosper. Those who do not, fail.
"The consumers determine, in the last resort, not only the prices of the consumers' goods, but no less the prices of all factors of production. They determine the income of every member of the market economy."
Ludwig von Mises, Human Action, 1949
Esto significa que cuando una revolución tecnológica transforma la economía, la dirección del cambio no la decide la tecnología: la decide la gente. Los consumidores, con sus preferencias, sus deseos, sus insatisfacciones, redirigen los recursos hacia donde perciben mayor valor. Si la IA abarata la traducción, los consumidores redirigirán el dinero ahorrado hacia otras necesidades. Si abarata el diseño gráfico, el dinero fluirá hacia donde haya mayor escasez percibida. No podemos predecir hacia dónde exactamente, pero podemos estar seguros de que fluirá, porque los deseos humanos no tienen techo.
Aplicado a la inteligencia artificial, este principio es al mismo tiempo tranquilizador y humilde. Tranquilizador porque nos asegura que la economía se reorganizará siguiendo los deseos de las personas, no los dictados de una máquina. Si los consumidores valoran la atención humana por encima de la eficiencia artificial, la economía se reorientará hacia la atención humana. Si valoran la autenticidad, la economía producirá autenticidad. La IA es una herramienta poderosísima, pero el rumbo lo marcan los deseos humanos. Y humilde porque nos recuerda que no sabemos, ni podemos saber, cómo reaccionarán exactamente las sociedades humanas ante la IA. Podemos hacer suposiciones, analizar tendencias, trazar escenarios. Pero la dirección real la marcarán millones de decisiones individuales que nadie puede anticipar, porque el conocimiento de las preferencias futuras no existe todavía: se revelará a medida que las personas experimenten, elijan y descarten.
La libertad de elegir
Hay un aspecto de la soberanía del consumidor que merece atención especial, porque contiene una verdad que los debates sobre la IA suelen ignorar: la tecnología no obliga, habilita.
Cada avance tecnológico amplía el menú de opciones disponibles para el ser humano, pero no elimina las opciones anteriores. La existencia de la comida industrial no impide cultivar tu propio huerto. La existencia de Spotify no impide comprar vinilos ni asistir a conciertos acústicos. La existencia de la enseñanza online no impide buscar un tutor presencial. La persona es libre de elegir entre más opciones, y eso incluye la opción de rechazar la tecnología y vivir con menos. Comunidades enteras eligen conscientemente estilos de vida con menos tecnología, desde los Amish hasta los movimientos de simplicidad voluntaria, y esa elección es legítima y respetable.
Lo que la tecnología hace es expandir las posibilidades. Antes del automóvil, solo podías ir a pie, a caballo o en carruaje. Después, podías además conducir, pero seguías pudiendo caminar. Antes de la IA, solo podías traducir pagando a un profesional o aprendiendo el idioma. Después, puedes además usar una máquina, pero sigues pudiendo contratar a un humano si lo prefieres.
Esta libertad de elección es la válvula de seguridad del sistema. Si la sociedad decide que en ciertos ámbitos prefiere lo humano a lo artificial, la economía se adaptará a esa preferencia, porque la economía de mercado no es otra cosa que un mecanismo para satisfacer lo que la gente elige. La IA no impone un futuro determinado. Ofrece posibilidades. Y es la humanidad, con sus decisiones, quien construye el futuro real.
La riqueza verdadera no es tener más cosas: es tener más opciones. Cuanto mayor es la capacidad tecnológica de una sociedad, mayor es su libertad real. Puedes elegir entre más alternativas, acceder a bienes de mayor valor, tomar decisiones más informadas. Y puedes incluso elegir volver a estados anteriores si así lo deseas. La tecnología no cierra puertas. Abre nuevas sin cerrar las antiguas.
La producción se alarga, no se elimina
Queda una pieza más en nuestro edificio teórico. Eugen von Böhm-Bawerk, maestro de Mises, observó algo sobre la producción que resulta especialmente iluminador.
Böhm-Bawerk notó que la producción más eficiente no es la más directa, sino la más indirecta. Un pescador que atrapa peces con las manos obtiene poco. Si dedica tiempo a fabricar una caña, pescará más. Si fabrica una red, aún más. Si construye un barco, incomparablemente más. Cada paso intermedio, cada bien de capital, aumenta la productividad final, pero requiere una inversión previa de tiempo y recursos.
A medida que una economía se desarrolla, su estructura productiva se alarga: hay más etapas intermedias entre la materia prima y el producto que llega al consumidor. Más capital invertido, más especialización, más coordinación. Esto significa que cuando la tecnología abarata los bienes de consumo final, no elimina la actividad económica: la desplaza hacia los bienes de orden superior que hacen posible esa producción.
Con la inteligencia artificial, este patrón se lleva al extremo. El texto generado, la imagen, el código, la traducción: todos estos bienes de consumo final se abaratan radicalmente. Pero la cadena productiva que hace posible esa abundancia es formidable: investigadores que diseñan algoritmos, ingenieros que construyen chips de última generación, centros de datos que consumen gigavatios de energía, cables submarinos que transmiten datos, minas de tierras raras que extraen los minerales necesarios, universidades que forman a los profesionales. Entrenar un modelo de inteligencia artificial de frontera cuesta cientos de millones de dólares. La inversión global en infraestructura de IA supera los doscientos mil millones anuales. Cuanto más barato es el producto final que la IA genera, más larga y costosa es la cadena productiva que lo sustenta. La economía no se contrae: se desplaza hacia capas más profundas de la estructura productiva.
Los límites del conocimiento humano
Hay una última reflexión que completa nuestro marco y que tiene que ver con los límites del desarrollo tecnológico mismo.
Cuando se habla de inteligencia artificial, es fácil dejarse llevar por la sensación de que la tecnología está a punto de resolverlo todo. Que si las máquinas piensan, pronto no quedará problema sin solución. Pero la realidad es que la humanidad, incluso potenciada por la IA, se encuentra todavía extraordinariamente lejos de resolver algunos de los problemas más fundamentales de la ciencia y la ingeniería.
Llevamos más de setenta años intentando lograr la fusión nuclear comercial, con decenas de miles de millones de dólares invertidos, y seguimos sin un reactor viable. No tenemos cura para el Alzheimer, el Parkinson ni la mayoría de enfermedades neurodegenerativas; los ensayos clínicos fracasan en más del noventa y nueve por ciento de los casos. No comprendemos qué es la consciencia ni por qué la actividad neuronal genera experiencia subjetiva. No podemos predecir el clima con fiabilidad más allá de dos semanas porque la meteorología es un sistema matemáticamente caótico. No sabemos qué es el noventa y cinco por ciento del universo. No podemos diseñar materiales con propiedades exactas a la carta. La colonización espacial sigue siendo un desafío titánico.
Y lo más revelador de todo: cada descubrimiento científico, en lugar de acercar el horizonte del conocimiento completo, lo aleja. Cuando descubrimos el átomo, descubrimos que contenía partículas menores. Cuando descubrimos esas partículas, encontramos quarks. Cuando secuenciamos el genoma humano, descubrimos la epigenética, el microbioma, las interacciones gen-ambiente: capas de complejidad que ni sospechábamos. La frontera del conocimiento no es estática. Retrocede a medida que avanzamos hacia ella.
La inteligencia artificial será sin duda un factor exponencial en la aceleración de la investigación y el desarrollo. Permitirá analizar cantidades de datos que ningún equipo humano podría procesar, descubrir patrones ocultos, simular escenarios, acelerar ciclos de experimentación. Pero no cambia el hecho fundamental de que el desarrollo tecnológico de la humanidad tiene por delante un camino inmenso, probablemente inabarcable. Cada problema resuelto revela problemas nuevos. Cada frontera conquistada descubre un horizonte más amplio. Y mientras haya problemas por resolver, habrá demanda de recursos, de investigación, de trabajo, de inversión. Habrá escasez de soluciones. Y donde hay escasez, como nos enseñó Menger, hay economía.
Esto no es un argumento contra la IA. Es un argumento a favor de ella. La IA no nos acerca al final del camino tecnológico: nos permite caminar más rápido por un camino que, por lo que sabemos, es infinito. Y eso es una oportunidad extraordinaria, no un motivo de pánico.
This means that when a technological revolution transforms the economy, the direction of change is not decided by technology: it is decided by people. Consumers, with their preferences, their desires, their dissatisfactions, redirect resources toward where they perceive greater value. If AI makes translation cheaper, consumers will redirect the money saved toward other needs. If it makes graphic design cheaper, the money will flow toward wherever there is greater perceived scarcity. We cannot predict exactly where, but we can be sure it will flow, because human desires have no ceiling.
Applied to artificial intelligence, this principle is at once reassuring and humbling. Reassuring because it assures us that the economy will reorganize itself following the desires of people, not the dictates of a machine. If consumers value human attention above artificial efficiency, the economy will reorient itself toward human attention. If they value authenticity, the economy will produce authenticity. AI is an immensely powerful tool, but the course is set by human desires. And humbling because it reminds us that we do not know, nor can we know, exactly how human societies will react to AI. We can make assumptions, analyze trends, sketch scenarios. But the real direction will be set by millions of individual decisions that no one can anticipate, because the knowledge of future preferences does not yet exist: it will be revealed as people experiment, choose, and discard.
The Freedom to Choose
There is an aspect of consumer sovereignty that deserves special attention, because it contains a truth that debates about AI tend to ignore: technology does not compel; it enables.
Every technological advance expands the menu of options available to the human being, but it does not eliminate previous options. The existence of industrial food does not prevent you from cultivating your own garden. The existence of Spotify does not prevent you from buying vinyl records or attending acoustic concerts. The existence of online education does not prevent you from seeking an in-person tutor. The person is free to choose among more options, and that includes the option to reject technology and live with less. Entire communities consciously choose lifestyles with less technology, from the Amish to voluntary simplicity movements, and that choice is legitimate and respectable.
What technology does is expand the possibilities. Before the automobile, you could only go on foot, on horseback, or by carriage. Afterward, you could also drive, but you could still walk. Before AI, you could only translate by paying a professional or learning the language. Afterward, you can also use a machine, but you can still hire a human if you prefer.
This freedom of choice is the system's safety valve. If society decides that in certain domains it prefers the human to the artificial, the economy will adapt to that preference, because the market economy is nothing other than a mechanism for satisfying what people choose. AI does not impose a predetermined future. It offers possibilities. And it is humanity, with its decisions, that builds the real future.
True wealth is not having more things: it is having more options. The greater a society's technological capacity, the greater its real freedom. You can choose among more alternatives, access goods of greater value, make more informed decisions. And you can even choose to return to previous states if you so wish. Technology does not close doors. It opens new ones without closing the old.
Production Lengthens; It Is Not Eliminated
One more piece remains in our theoretical edifice. Eugen von Bohm-Bawerk, Mises's teacher, observed something about production that proves especially illuminating.
Bohm-Bawerk noted that the most efficient production is not the most direct, but the most roundabout. A fisherman who catches fish with his bare hands obtains little. If he spends time making a rod, he will catch more. If he makes a net, even more. If he builds a boat, incomparably more. Each intermediate step, each capital good, increases final productivity, but requires a prior investment of time and resources.
As an economy develops, its productive structure lengthens: there are more intermediate stages between the raw material and the product that reaches the consumer. More capital invested, more specialization, more coordination. This means that when technology makes final consumer goods cheaper, it does not eliminate economic activity: it displaces it toward the higher-order goods that make that production possible.
With artificial intelligence, this pattern is taken to the extreme. Generated text, images, code, translation: all these final consumer goods become radically cheaper. But the productive chain that makes that abundance possible is formidable: researchers designing algorithms, engineers building cutting-edge chips, data centers consuming gigawatts of energy, submarine cables transmitting data, rare-earth mines extracting the necessary minerals, universities training the professionals. Training a frontier artificial intelligence model costs hundreds of millions of dollars. Global investment in AI infrastructure exceeds two hundred billion annually. The cheaper the final product that AI generates, the longer and more costly the productive chain that supports it. The economy does not contract: it shifts toward deeper layers of the productive structure.
The Limits of Human Knowledge
There is one final reflection that completes our framework, and it has to do with the limits of technological development itself.
When people talk about artificial intelligence, it is easy to be swept away by the feeling that technology is about to solve everything. That if machines think, soon no problem will remain unsolved. But the reality is that humanity, even empowered by AI, is still extraordinarily far from solving some of the most fundamental problems of science and engineering.
We have been trying for more than seventy years to achieve commercial nuclear fusion, with tens of billions of dollars invested, and we still do not have a viable reactor. We have no cure for Alzheimer's, Parkinson's, or most neurodegenerative diseases; clinical trials fail in more than ninety-nine percent of cases. We do not understand what consciousness is or why neural activity generates subjective experience. We cannot predict the weather reliably more than two weeks out because meteorology is a mathematically chaotic system. We do not know what ninety-five percent of the universe is. We cannot design materials with exact properties to order. Space colonization remains a titanic challenge.
And most revealing of all: every scientific discovery, rather than bringing the horizon of complete knowledge closer, pushes it further away. When we discovered the atom, we discovered it contained smaller particles. When we discovered those particles, we found quarks. When we sequenced the human genome, we discovered epigenetics, the microbiome, gene-environment interactions: layers of complexity we had not even suspected. The frontier of knowledge is not static. It recedes as we advance toward it.
Artificial intelligence will undoubtedly be an exponential factor in the acceleration of research and development. It will allow the analysis of quantities of data that no human team could process, discover hidden patterns, simulate scenarios, accelerate experimentation cycles. But it does not change the fundamental fact that humanity's technological development has before it an immense path, probably an inexhaustible one. Every problem solved reveals new problems. Every frontier conquered reveals a broader horizon. And as long as there are problems to solve, there will be demand for resources, for research, for work, for investment. There will be scarcity of solutions. And where there is scarcity, as Menger taught us, there is economy.
This is not an argument against AI. It is an argument in its favor. AI does not bring us closer to the end of the technological road: it allows us to walk faster along a road that, as far as we know, is infinite. And that is an extraordinary opportunity, not a cause for panic.
04 La IA ante los principios inmutables AI Against the Immutable Principles
Ahora que tenemos el marco teórico completo, podemos hacer lo que los titulares alarmistas nunca hacen: analizar las preocupaciones sobre la inteligencia artificial con rigor, una por una, a la luz de principios que no han fallado en doscientos años.
«La IA hará todo el trabajo. No quedará nada para los humanos.»
Esta es la preocupación reina, la más extendida y la que más angustia genera. Y se desmonta directamente con Mises.
Para que no quedara «nada para los humanos», necesitaríamos que la acción humana cesara. Y la acción humana cesa solo cuando desaparece la insatisfacción, la imaginación o la capacidad de actuar. Hemos visto que ninguna de las tres puede desaparecer: la insatisfacción es consustancial a la naturaleza humana, la imaginación no tiene techo y la IA amplifica la capacidad de actuar en lugar de reducirla.
Pero hay un argumento aún más concreto. David Ricardo, el gran economista clásico, demostró en 1817 algo extraordinariamente contraintuitivo: incluso cuando un agente es superior a otro en absolutamente todas las tareas, sigue habiendo beneficio en la especialización y el intercambio. Es el principio de la ventaja comparativa.
Pensemos en un cirujano que además es el mejor cocinero de su hospital. ¿Debería cocinar su propia comida? Evidentemente no, porque cada hora que pasa cocinando es una hora que no pasa operando, donde su valor es incomparablemente mayor. Le conviene que otra persona cocine, aunque esa persona sea peor cocinera que él.
Con la IA ocurrirá lo mismo. Supongamos que es superior al humano en investigación médica (donde su ventaja es de 100 a 1) y también en atención emocional al paciente (donde su ventaja es de 2 a 1). Es económicamente absurdo que la IA dedique sus recursos a atención emocional, porque cada unidad de tiempo que dedica ahí deja de producir investigación donde su ventaja es cincuenta veces mayor. Es más eficiente que se concentre en lo que hace extraordinariamente mejor y deje a los humanos lo que hace solo ligeramente mejor. No porque los humanos sean superiores, sino porque el coste de oportunidad de no hacerlo así es enorme.
«Si la IA produce todo casi gratis, la economía se derrumbará.»
Aquí es donde Menger y la lección de la historia son devastadores.
En 1800, una familia europea gastaba el ochenta por ciento de su ingreso en comida. Hoy gasta entre el diez y el quince por ciento. La comida se abarató un ochenta y cinco por ciento en términos relativos. Si la lógica de «precio casi gratis implica fin de la industria» fuera cierta, la industria alimentaria debería haber colapsado. En cambio, es hoy incomparablemente más grande, más diversa y más sofisticada que en cualquier momento de la historia. Restaurantes de alta cocina, comida rápida, delivery, alimentación orgánica, suplementos nutricionales, turismo gastronómico: industrias que mueven billones de dólares y que nacieron porque la comida básica se abarató.
El principio es claro: cuando el precio de un bien baja, la demanda no se mantiene constante. Se expande a dimensiones que antes eran inaccesibles. En 1800 la demanda de comida era «no morir de hambre». Cuando esa necesidad se resolvió de forma barata, emergieron necesidades nuevas: comer bien, comer variado, comer cómodo, comer sano, comer experiencial. Cada una creó mercados nuevos.
Con la IA veremos exactamente lo mismo. La traducción casi gratuita no destruirá la economía: abrirá mercados internacionales a millones de pequeñas empresas que nunca pudieron costear traductores, y cada una necesitará logística, atención al cliente, marketing local, adaptación cultural. El software casi gratuito permitirá que millones de emprendedores lancen proyectos que antes eran inviables. El dinero que los consumidores ahorren no desaparecerá: se redirigirá hacia nuevas necesidades que hoy ni podemos nombrar.
«Pero esta vez es diferente. La IA piensa. Las máquinas anteriores solo ejecutaban.»
Es el argumento más sofisticado y merece una respuesta cuidadosa. Es cierto que la IA representa un salto cualitativo respecto a las máquinas anteriores. No solo automatiza tareas mecánicas: procesa lenguaje, genera contenido, analiza patrones complejos, toma decisiones en entornos inciertos. Es legítimo preguntarse si esta diferencia invalida las lecciones de las revoluciones anteriores.
La respuesta, desde la Escuela Austriaca, es que no. Y la razón no está en las limitaciones de la IA, sino en la naturaleza de la economía.
Recordemos a Hayek: el problema económico no es un problema de cálculo ni de inteligencia. Es un problema de conocimiento disperso. Aunque la IA sea capaz de pensar, calcular y crear, no puede agregar el conocimiento tácito y contextual de millones de personas que están decidiendo, en este preciso instante, qué valoran, qué necesitan, qué están dispuestas a intercambiar. Ese conocimiento no está en ninguna base de datos. No es digitalizable.
Además, recordemos a Mises: la economía no está impulsada por la inteligencia, sino por la acción humana. La acción humana brota de la insatisfacción subjetiva, de la imaginación personal, de la valoración individual de lo que constituye una vida mejor. La IA puede ser más inteligente que cualquier humano, pero no puede desear por nosotros. No puede decidir que nuestra vida sería mejor con más tiempo en la naturaleza, o con una relación más profunda con nuestros hijos, o con una carrera profesional diferente. Esas valoraciones son irreductiblemente humanas, y son ellas las que dirigen la economía.
Y recordemos a Menger: la diferencia cualitativa de la IA no cambia el mecanismo fundamental. La tecnología abarata bienes existentes, los consumidores redirigen los recursos liberados hacia nuevas necesidades, emergen nuevas escaseces y nuevos mercados. Este mecanismo no depende de si la máquina es mecánica, eléctrica o inteligente. Depende de la naturaleza de los deseos humanos y de la escasez, y ninguna de las dos cambia porque la IA piense.
«La mayoría de la gente quedará sin trabajo.»
Esta preocupación asume que el número de trabajos posibles es limitado y que la IA los absorberá uno a uno hasta no dejar nada. Es, de nuevo, la falacia del pastel fijo vestida con ropa nueva.
Recordemos el mecanismo: cuando la IA abarata la producción de un bien, el dinero que los consumidores ahorran se redirige hacia otros bienes y servicios. Esos otros bienes y servicios requieren producción, y esa producción requiere trabajo. El trabajo no desaparece: migra. Como migró de la agricultura a la industria, de la industria a los servicios, de los servicios a la economía digital.
Y recordemos también la soberanía del consumidor. Los consumidores, como seres humanos con deseos infinitos e insatisfacción perpetua, siempre encuentran nuevas formas de gastar sus recursos. Lo que gastarán con el dinero ahorrado gracias a la IA no lo sabemos hoy, como no sabíamos en 1995 que la gente pagaría por ver a desconocidos jugar a videojuegos en directo o que existiría un mercado multimillonario para la comida a domicilio. Pero la historia nos asegura que lo gastarán, porque la alternativa, la satisfacción plena, el fin de todo deseo, es algo que no ha ocurrido jamás.
«El proceso será demasiado rápido. La gente no tendrá tiempo de adaptarse.»
Es una preocupación legítima y merece una respuesta honesta. Es cierto que la velocidad de transformación de la IA es mayor que la de revoluciones anteriores.
Pero la velocidad de destrucción es también la velocidad de creación. Si la IA abarata servicios rápidamente, el poder adquisitivo se libera rápidamente. Si el dinero fluye rápidamente hacia nuevas necesidades, los nuevos mercados emergen rápidamente. La transición es más intensa, pero no es unidireccional: destrucción y creación ocurren simultáneamente, no secuencialmente.
Además, la IA misma es una herramienta de adaptación. El trabajador que necesita reciclarse tiene hoy acceso a educación personalizada casi gratuita, a asistentes que le enseñan nuevas habilidades a su ritmo, a herramientas que amplifican su productividad en campos que desconocía. La misma tecnología que transforma los empleos proporciona los medios para adaptarse a esa transformación. Ninguna revolución anterior ofreció eso.
«La riqueza se concentrará en quienes posean la IA.»
Esta preocupación confunde la distribución de la riqueza con su creación. Son problemas distintos. Es posible, y es un debate legítimo, que la transición genere desigualdades temporales o que ciertos actores acumulen ventajas desproporcionadas. Pero eso no invalida el principio de que la riqueza total aumenta.
La historia ofrece un precedente claro. Durante la Revolución Industrial, los dueños de las primeras fábricas acumularon fortunas enormes. La desigualdad creció durante décadas. Pero el resultado a largo plazo fue que la riqueza total se multiplicó de tal forma que incluso los trabajadores de 1900 vivían mejor que los artesanos prósperos de 1750.
Además, la IA tiene una característica que las revoluciones anteriores no tuvieron: su poder se democratiza a una velocidad extraordinaria. El acceso a modelos de lenguaje, a herramientas de generación de código, de diseño, de análisis, está disponible para cualquier persona con una conexión a Internet. Un emprendedor en un país en desarrollo puede acceder hoy a capacidades que hace cinco años estaban reservadas a grandes corporaciones. Esta democratización del poder productivo no tiene precedente y actúa como contrapeso natural a la concentración.
«Una IA superinteligente podría planificar toda la economía.»
Mises respondió a este argumento en 1920, treinta años antes de que existieran los primeros ordenadores. La respuesta es tan vigente hoy como entonces.
El problema de la planificación central no es un problema de capacidad de cálculo. Es un problema de información. Los precios de mercado son señales que condensan las valoraciones subjetivas de millones de personas. Destruir el mercado para sustituirlo por un planificador, por inteligente que sea, no es un avance: es destruir el único mecanismo que genera esa información.
Una IA puede calcular a velocidades inimaginables, pero no puede saber qué valora cada persona en cada momento sin un sistema de precios que lo revele. Es como tener el procesador más potente del mundo pero sin datos que procesar: la potencia es irrelevante si la información no existe en una forma accesible.
Now that we have the complete theoretical framework, we can do what alarmist headlines never do: analyze concerns about artificial intelligence rigorously, one by one, in the light of principles that have not failed in two hundred years.
"AI will do all the work. There will be nothing left for humans."
This is the reigning concern, the most widespread and the one that generates the most anxiety. And it is dismantled directly by Mises.
For there to be "nothing left for humans," human action would need to cease. And human action ceases only when dissatisfaction, imagination, or the capacity to act disappears. We have seen that none of the three can disappear: dissatisfaction is inherent to human nature, imagination has no ceiling, and AI amplifies the capacity to act instead of reducing it.
But there is an even more concrete argument. David Ricardo, the great classical economist, demonstrated in 1817 something extraordinarily counterintuitive: even when one agent is superior to another in absolutely every task, there is still benefit in specialization and exchange. This is the principle of comparative advantage.
Think of a surgeon who also happens to be the best cook in the hospital. Should they cook their own food? Obviously not, because every hour spent cooking is an hour not spent operating, where their value is incomparably greater. It is better for someone else to cook, even if that person is a worse cook than the surgeon.
The same will happen with AI. Suppose it is superior to humans in medical research (where its advantage is 100 to 1) and also in emotional patient care (where its advantage is 2 to 1). It is economically absurd for the AI to dedicate its resources to emotional care, because every unit of time spent there is a unit not producing research where its advantage is fifty times greater. It is more efficient for it to concentrate on what it does extraordinarily better and leave to humans what it does only slightly better. Not because humans are superior, but because the opportunity cost of not doing so is enormous.
"If AI produces everything almost for free, the economy will collapse."
This is where Menger and the lesson of history are devastating.
In 1800, a European family spent eighty percent of its income on food. Today it spends between ten and fifteen percent. Food became eighty-five percent cheaper in relative terms. If the logic of "near-free price implies the end of the industry" were true, the food industry should have collapsed. Instead, it is today incomparably larger, more diverse, and more sophisticated than at any point in history. Fine dining restaurants, fast food, delivery, organic food, nutritional supplements, gastronomic tourism: industries that move trillions of dollars and that were born because basic food became cheap.
The principle is clear: when the price of a good falls, demand does not remain constant. It expands to dimensions that were previously inaccessible. In 1800, the demand for food was "not dying of hunger." When that need was solved cheaply, new needs emerged: eating well, eating varied, eating conveniently, eating healthily, eating experientially. Each created new markets.
With AI we will see exactly the same thing. Near-free translation will not destroy the economy: it will open international markets to millions of small businesses that could never afford translators, and each will need logistics, customer service, local marketing, cultural adaptation. Near-free software will allow millions of entrepreneurs to launch projects that were previously unfeasible. The money consumers save will not disappear: it will be redirected toward new needs that today we cannot even name.
"But this time is different. AI thinks. Previous machines only executed."
This is the most sophisticated argument and it deserves a careful answer. It is true that AI represents a qualitative leap compared to previous machines. It does not only automate mechanical tasks: it processes language, generates content, analyzes complex patterns, makes decisions in uncertain environments. It is legitimate to ask whether this difference invalidates the lessons of previous revolutions.
The answer, from the Austrian School, is that it does not. And the reason lies not in AI's limitations, but in the nature of the economy.
Let us recall Hayek: the economic problem is not a problem of calculation or of intelligence. It is a problem of dispersed knowledge. Even though AI is capable of thinking, calculating, and creating, it cannot aggregate the tacit and contextual knowledge of millions of people who are deciding, at this very instant, what they value, what they need, what they are willing to exchange. That knowledge is not in any database. It is not digitizable.
Furthermore, let us recall Mises: the economy is not driven by intelligence, but by human action. Human action springs from subjective dissatisfaction, from personal imagination, from the individual assessment of what constitutes a better life. AI may be more intelligent than any human, but it cannot desire for us. It cannot decide that our life would be better with more time in nature, or with a deeper relationship with our children, or with a different career. Those valuations are irreducibly human, and they are the ones that direct the economy.
And let us recall Menger: AI's qualitative difference does not change the fundamental mechanism. Technology makes existing goods cheaper, consumers redirect freed resources toward new needs, new scarcities and new markets emerge. This mechanism does not depend on whether the machine is mechanical, electrical, or intelligent. It depends on the nature of human desires and of scarcity, and neither of those changes because AI thinks.
"Most people will be left without work."
This concern assumes that the number of possible jobs is limited and that AI will absorb them one by one until nothing is left. It is, once again, the fixed pie fallacy dressed in new clothes.
Let us recall the mechanism: when AI makes the production of a good cheaper, the money consumers save is redirected toward other goods and services. Those other goods and services require production, and that production requires work. Work does not disappear: it migrates. As it migrated from agriculture to industry, from industry to services, from services to the digital economy.
And let us also recall consumer sovereignty. Consumers, as human beings with infinite desires and perpetual dissatisfaction, always find new ways to spend their resources. What they will spend with the money saved thanks to AI we do not know today, just as we did not know in 1995 that people would pay to watch strangers play video games live, or that there would be a multibillion-dollar market for food delivery. But history assures us that they will spend it, because the alternative — full satisfaction, the end of all desire — is something that has never occurred.
"The process will be too fast. People will not have time to adapt."
This is a legitimate concern and it deserves an honest answer. It is true that AI's speed of transformation is greater than that of previous revolutions.
But the speed of destruction is also the speed of creation. If AI makes services cheaper rapidly, purchasing power is freed rapidly. If money flows rapidly toward new needs, new markets emerge rapidly. The transition is more intense, but it is not unidirectional: destruction and creation occur simultaneously, not sequentially.
Moreover, AI itself is a tool for adaptation. The worker who needs to retrain today has access to nearly free personalized education, to assistants that teach new skills at their own pace, to tools that amplify their productivity in fields they previously knew nothing about. The same technology that transforms jobs provides the means to adapt to that transformation. No previous revolution offered that.
"Wealth will be concentrated in those who own the AI."
This concern conflates the distribution of wealth with its creation. They are distinct problems. It is possible, and it is a legitimate debate, that the transition will generate temporary inequalities or that certain actors will accumulate disproportionate advantages. But that does not invalidate the principle that total wealth increases.
History offers a clear precedent. During the Industrial Revolution, the owners of the first factories accumulated enormous fortunes. Inequality grew for decades. But the long-term result was that total wealth multiplied to such an extent that even the workers of 1900 lived better than the prosperous artisans of 1750.
Furthermore, AI has a characteristic that previous revolutions did not: its power is democratized at an extraordinary speed. Access to language models, to code generation tools, design tools, analysis tools, is available to anyone with an Internet connection. An entrepreneur in a developing country can today access capabilities that five years ago were reserved for large corporations. This democratization of productive power is unprecedented and acts as a natural counterweight to concentration.
"A superintelligent AI could plan the entire economy."
Mises answered this argument in 1920, thirty years before the first computers existed. The answer is as valid today as it was then.
The problem of central planning is not a problem of computing power. It is a problem of information. Market prices are signals that condense the subjective valuations of millions of people. Destroying the market to replace it with a planner, however intelligent, is not an advance: it is destroying the only mechanism that generates that information.
An AI can calculate at unimaginable speeds, but it cannot know what each person values at each moment without a price system to reveal it. It is like having the most powerful processor in the world but with no data to process: the power is irrelevant if the information does not exist in an accessible form.
Lo que la historia y la teoría nos aseguran
What History and Theory Assure Us
Hemos recorrido un largo camino. Desde los luditas rompiendo telares en la Inglaterra de 1811 hasta los titulares que hoy anuncian que la inteligencia artificial acabará con el trabajo humano. Desde los fundamentos de la acción humana hasta los límites del conocimiento. Desde la escasez que no desaparece hasta el consumidor que guía la economía con sus decisiones libres.
Todo apunta en la misma dirección.
La inteligencia artificial transformará la economía profundamente. Destruirá empleos, industrias y modelos de negocio. Eso es real y merece atención seria. Pero no destruirá la actividad económica, porque la actividad económica no depende de qué herramientas usamos. Depende de que los seres humanos estemos insatisfechos, de que podamos imaginar algo mejor, de que tengamos medios para buscarlo, de que los recursos sean escasos, de que haya que elegir, de que la propiedad exista, de que los precios transmitan información, de que los consumidores dirijan la economía con sus deseos siempre crecientes, de que la libertad de elegir se mantenga.
Ninguna de estas condiciones desaparece con la IA. Todas se intensifican.
Cada revolución tecnológica erosiona la economía tal como la conocemos. Y cada vez, lo que emerge en su lugar es más grande, más diverso y más rico que lo anterior. La economía transiciona, se reconfigura, se eleva a un nivel superior donde la riqueza material es mayor, los deseos y necesidades son más sofisticados, y el horizonte tecnológico sigue siendo inmenso. No hay razón para pensar que la inteligencia artificial será la excepción. Hay siglos de evidencia y principios económicos inmutables para pensar que será la confirmación más espectacular de la regla.
La economía no llega nunca a un estado de reposo porque el ser humano no llega nunca a un estado de satisfacción plena. Mientras haya un solo deseo insatisfecho en una sola mente humana, habrá acción. Y donde hay acción, hay economía. La IA no apaga ese motor. Le da más combustible.
We have traveled a long road. From the Luddites breaking looms in England in 1811 to the headlines that today announce that artificial intelligence will end human work. From the foundations of human action to the limits of knowledge. From the scarcity that does not disappear to the consumer who guides the economy with free decisions.
Everything points in the same direction.
Artificial intelligence will transform the economy profoundly. It will destroy jobs, industries, and business models. That is real and deserves serious attention. But it will not destroy economic activity, because economic activity does not depend on what tools we use. It depends on human beings being dissatisfied, on our ability to imagine something better, on having the means to pursue it, on resources being scarce, on the necessity of choosing, on the existence of property, on prices transmitting information, on consumers directing the economy with their ever-growing desires, on the freedom to choose being preserved.
None of these conditions disappears with AI. All of them intensify.
Every technological revolution erodes the economy as we know it. And every time, what emerges in its place is larger, more diverse, and richer than what came before. The economy transitions, reconfigures, elevates itself to a higher level where material wealth is greater, desires and needs are more sophisticated, and the technological horizon remains immense. There is no reason to think that artificial intelligence will be the exception. There are centuries of evidence and immutable economic principles to believe it will be the most spectacular confirmation of the rule.
The economy never reaches a state of rest because the human being never reaches a state of full satisfaction. As long as there is a single unsatisfied desire in a single human mind, there will be action. And where there is action, there is economy. AI does not shut off that engine. It gives it more fuel.