For years, the @Dr_CSWright account was managed by others in my name, including throughout the court case. That arrangement has now ended. I have taken direct ownership and control of the account and will be posting personally from this point forward.
I accept no responsibility for anything posted here before I assumed control. I am not deleting the historical material; the past should occasionally be preserved, if only to demonstrate what happens when other people are entrusted with one’s voice. From now on, however, the words are mine.
@Dr_CSWright is now in my control.
https://x.com/i/status/2078514675881316764
https://t.me/S_Tominaga/5457
I accept no responsibility for anything posted here before I assumed control. I am not deleting the historical material; the past should occasionally be preserved, if only to demonstrate what happens when other people are entrusted with one’s voice. From now on, however, the words are mine.
@Dr_CSWright is now in my control.
https://x.com/i/status/2078514675881316764
https://t.me/S_Tominaga/5457
X (formerly Twitter)
S Tominaga (Aka Dr Craig Wright) (@CsTominaga) on X
For years, the @Dr_CSWright account was managed by others in my name, including throughout the court case. That arrangement has now ended. I have taken direct ownership and control of the account …
Durante años, la cuenta @Dr_CSWright fue administrada por otras personas en mi nombre, incluso durante el proceso judicial. Dicho acuerdo ha finalizado. Ahora soy el propietario y responsable directo de la cuenta y publicaré personalmente a partir de ahora.
No me hago responsable de nada publicado aquí antes de que asumiera el control. No voy a borrar el material anterior; el pasado a veces debe conservarse, aunque solo sea para demostrar lo que sucede cuando se confía la voz de uno a otras personas. Sin embargo, a partir de ahora, las palabras son mías.
@Dr_CSWright está ahora bajo mi control.
https://x.com/i/status/2078514675881316764
https://t.me/S_Tominaga/5457
No me hago responsable de nada publicado aquí antes de que asumiera el control. No voy a borrar el material anterior; el pasado a veces debe conservarse, aunque solo sea para demostrar lo que sucede cuando se confía la voz de uno a otras personas. Sin embargo, a partir de ahora, las palabras son mías.
@Dr_CSWright está ahora bajo mi control.
https://x.com/i/status/2078514675881316764
https://t.me/S_Tominaga/5457
X (formerly Twitter)
S Tominaga (Aka Dr Craig Wright) (@CsTominaga) on X
For years, the @Dr_CSWright account was managed by others in my name, including throughout the court case. That arrangement has now ended. I have taken direct ownership and control of the account …
Everyone is asking whether AI is safe. Almost no one is asking the question that actually determines the answer: when an AI system exercises power over you, who owes you a duty — and what is that duty?
Right now the answer is nobody, and nothing. That is the real crisis, and it is a legal one.
We have built our entire AI accountability conversation on the wrong word. We say "responsible AI," "trustworthy AI," "aligned AI" — adjectives describing the system. But responsibility, trust and duty are not properties of software. They are relationships between people. A system cannot owe you anything. The people who exercise control through it can. The whole game is locating them.
That location is exactly what modern governance is engineered to prevent. "The model decided." "The algorithm flagged you." "The system recommended it." This grammar is not innocent. It is a machine for dissolving authorship — for making sure that when power lands on a person, no human hand can be found on it. A denied loan, a rejected claim, a flagged account, a sentencing score: power was exercised, an interest was harmed, and the org's first move is to point at the model as if the model were the author.
The law has a name for what should happen when someone exercises discretionary power over another's interests while that person is vulnerable to the exercise of it. It is called a fiduciary duty. Loyalty. Good faith. No self-dealing. It does not require a contract. It does not require a corporate form. It attaches to the exercise of power itself. And it is precisely the concept missing from AI governance.
In my work I develop a framework I call the Attribution Stack — four layers of control for finding responsibility in systems designed to diffuse it: formal command, structural inducement, default power, and veto authority. I built it to answer a blockchain question — who actually governs a network that claims to be governed by no one — but the structure is general, and AI is where it bites hardest.
Run any AI deployment through the stack and the diffusion collapses.
Formal command: who set the objective the system optimises? Someone chose the target. A model does not choose what counts as success; a person does, and that choice is an act of authorship.
Structural inducement: who shaped the incentives — the engagement metric, the cost function, the deployment pressure — that made the harmful behaviour the rational output rather than an accident?
Default power: who controls what happens when no one intervenes? The default is a decision. A system that denies by default and requires the vulnerable party to appeal has been designed by someone, and that someone is answerable for the design.
Veto authority: who could have stopped it and did not? The power to halt a deployment, override an output, or pull a model is control, whether or not it is exercised.
At every layer there is a human with a name. The diffusion is a story, not a fact. And once you can attribute the control, you can attach the duty — because the duty follows the power, not the org chart.
This is why "human in the loop" is not the answer people think it is. A person clicking approve on outputs they cannot inspect, under time pressure, with no authority to refuse, is not a decision-maker. They are a liability sponge — placed there so the institution can say a human decided, while the human absorbs the blame for a system whose objective, incentives, defaults and off-switch were all controlled by someone else. A signature is not answerability. It is often the opposite: answerability's disguise.
So here is the claim, stated plainly. The entities that exercise discretionary control over consequential AI systems
1/2
Dr. Craig S Wright
Jul 18, 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
Right now the answer is nobody, and nothing. That is the real crisis, and it is a legal one.
We have built our entire AI accountability conversation on the wrong word. We say "responsible AI," "trustworthy AI," "aligned AI" — adjectives describing the system. But responsibility, trust and duty are not properties of software. They are relationships between people. A system cannot owe you anything. The people who exercise control through it can. The whole game is locating them.
That location is exactly what modern governance is engineered to prevent. "The model decided." "The algorithm flagged you." "The system recommended it." This grammar is not innocent. It is a machine for dissolving authorship — for making sure that when power lands on a person, no human hand can be found on it. A denied loan, a rejected claim, a flagged account, a sentencing score: power was exercised, an interest was harmed, and the org's first move is to point at the model as if the model were the author.
The law has a name for what should happen when someone exercises discretionary power over another's interests while that person is vulnerable to the exercise of it. It is called a fiduciary duty. Loyalty. Good faith. No self-dealing. It does not require a contract. It does not require a corporate form. It attaches to the exercise of power itself. And it is precisely the concept missing from AI governance.
In my work I develop a framework I call the Attribution Stack — four layers of control for finding responsibility in systems designed to diffuse it: formal command, structural inducement, default power, and veto authority. I built it to answer a blockchain question — who actually governs a network that claims to be governed by no one — but the structure is general, and AI is where it bites hardest.
Run any AI deployment through the stack and the diffusion collapses.
Formal command: who set the objective the system optimises? Someone chose the target. A model does not choose what counts as success; a person does, and that choice is an act of authorship.
Structural inducement: who shaped the incentives — the engagement metric, the cost function, the deployment pressure — that made the harmful behaviour the rational output rather than an accident?
Default power: who controls what happens when no one intervenes? The default is a decision. A system that denies by default and requires the vulnerable party to appeal has been designed by someone, and that someone is answerable for the design.
Veto authority: who could have stopped it and did not? The power to halt a deployment, override an output, or pull a model is control, whether or not it is exercised.
At every layer there is a human with a name. The diffusion is a story, not a fact. And once you can attribute the control, you can attach the duty — because the duty follows the power, not the org chart.
This is why "human in the loop" is not the answer people think it is. A person clicking approve on outputs they cannot inspect, under time pressure, with no authority to refuse, is not a decision-maker. They are a liability sponge — placed there so the institution can say a human decided, while the human absorbs the blame for a system whose objective, incentives, defaults and off-switch were all controlled by someone else. A signature is not answerability. It is often the opposite: answerability's disguise.
So here is the claim, stated plainly. The entities that exercise discretionary control over consequential AI systems
1/2
Dr. Craig S Wright
Jul 18, 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
— that set the objectives, shape the incentives, own the defaults and hold the veto — should be treated as fiduciaries to the people those systems act upon. Not metaphorically. Doctrinally. With duties of loyalty and care, a prohibition on self-dealing against the interests of those they govern, and remedies that survive the sentence "the model did it."
We do not need to invent this from nothing. Partnership law, fiduciary doctrine and property law already know how to find authorship behind a facade and attach obligation to power. What they have lacked is the willingness to look past the story a system tells about itself — the story that it is autonomous, decentralised, leaderless, just math.
It is never just math. Someone chose the objective. Someone owns the default. Someone holds the veto.
The defining legal challenge of this era is not regulating AI as if it were a force of nature. It is refusing the disappearance of the people who govern it — and imposing the duties that power has always attracted, on the humans who have simply learned to exercise it through systems that make them hard to see.
Attribute the control. Attach the duty.
2/2
Dr. Craig S Wright
Jul 18, 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
We do not need to invent this from nothing. Partnership law, fiduciary doctrine and property law already know how to find authorship behind a facade and attach obligation to power. What they have lacked is the willingness to look past the story a system tells about itself — the story that it is autonomous, decentralised, leaderless, just math.
It is never just math. Someone chose the objective. Someone owns the default. Someone holds the veto.
The defining legal challenge of this era is not regulating AI as if it were a force of nature. It is refusing the disappearance of the people who govern it — and imposing the duties that power has always attracted, on the humans who have simply learned to exercise it through systems that make them hard to see.
Attribute the control. Attach the duty.
2/2
Dr. Craig S Wright
Jul 18, 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
Todo el mundo se pregunta si la IA es segura. Casi nadie se plantea la pregunta que realmente determina la respuesta: cuando un sistema de IA ejerce poder sobre nosotros, ¿quién tiene una obligación con nosotros? ¿Y cuál es esa obligación?
Ahora mismo, la respuesta es nadie, y nada. Esa es la verdadera crisis, y es una crisis legal.
Hemos construido todo nuestro debate sobre la responsabilidad de la IA sobre la base de la palabra equivocada. Hablamos de "IA responsable", "IA confiable", "IA alineada": adjetivos que describen el sistema. Pero la responsabilidad, la confianza y la obligación no son propiedades del software. Son relaciones entre personas. Un sistema no puede debernos nada. Quienes ejercen el control a través de él sí. La clave está en identificarlos.
Esa identificación es precisamente lo que la gobernanza moderna busca evitar. "El modelo decidió". "El algoritmo te señaló". "El sistema lo recomendó". Esta gramática no es inocente. Es una herramienta para disolver la autoría, para asegurar que cuando el poder recaiga sobre una persona, no se pueda encontrar ninguna huella humana en ella. Un préstamo denegado, una reclamación rechazada, una cuenta marcada, una calificación de riesgo: se ejerció poder, se perjudicó un interés, y la primera reacción de la organización es culpar al modelo como si fuera el autor.
La ley define lo que debería ocurrir cuando alguien ejerce poder discrecional sobre los intereses de otra persona mientras esta es vulnerable a dicho ejercicio. Se denomina deber fiduciario. Lealtad. Buena fe. No hay conflicto de intereses. No requiere contrato. No requiere forma jurídica. Está ligado al ejercicio del poder en sí mismo. Y es precisamente el concepto que falta en la gobernanza de la IA.
En mi trabajo, desarrollo un marco que denomino Pila de Atribución: cuatro niveles de control para determinar la responsabilidad en sistemas diseñados para diluirla: mandato formal, inducción estructural, poder por defecto y derecho de veto. Lo construí para responder a una pregunta sobre blockchain: ¿quién gobierna realmente una red que afirma no estar gobernada por nadie? Sin embargo, la estructura es general, y es en la IA donde más se manifiesta.
Si se ejecuta cualquier implementación de IA a través de la pila tecnológica, la difusión se desmorona.
Comando formal: ¿quién estableció el objetivo del mejor sistema? Alguien eligió el objetivo. Un modelo no decide qué se considera éxito; lo hace una persona, y esa elección es un acto de autoría.
Incentivo estructural: ¿quién diseñó los incentivos —la métrica de participación, la función de costos, la presión de implementación— que convirtieron el comportamiento perjudicial en el resultado racional en lugar de un accidente?
Poder por defecto: ¿quién controla lo que sucede cuando nadie interviene? El valor por defecto es una decisión. Un sistema que niega por defecto y exige que la parte vulnerable apele ha sido diseñado por alguien, y esa persona es responsable de su diseño.
Autoridad de veto: ¿quién pudo haberlo detenido y no lo hizo? El poder de detener una implementación, anular un resultado o retirar un modelo es control, se ejerza o no.
Pero en cada capa hay una persona con un nombre. La difusión es una historia, no un hecho. Y una vez que se puede atribuir el control, se puede asignar la responsabilidad, porque la responsabilidad se deriva del poder, no del organigrama.
Por eso, la intervención humana no es la solución que muchos creen. Una persona que aprueba resultados que no puede inspeccionar, bajo presión de tiempo y sin autoridad para rechazarlos, no toma decisiones. Es una especie de esponja de responsabilidad, colocada ahí para que la institución pueda decir que un humano decidió, mientras que ese humano absorbe la culpa de un sistema cuyos objetivos, incentivos, configuraciones predeterminadas y mecanismo de desactivación estaban controlados por otra persona. Una firma no implica responsabilidad. A menudo es lo contrario: un disfraz de responsabilidad.
1/2
Dr. Craig S Wright
18 jul 2026
https://x.com/i/status/2078498382058852361
Ahora mismo, la respuesta es nadie, y nada. Esa es la verdadera crisis, y es una crisis legal.
Hemos construido todo nuestro debate sobre la responsabilidad de la IA sobre la base de la palabra equivocada. Hablamos de "IA responsable", "IA confiable", "IA alineada": adjetivos que describen el sistema. Pero la responsabilidad, la confianza y la obligación no son propiedades del software. Son relaciones entre personas. Un sistema no puede debernos nada. Quienes ejercen el control a través de él sí. La clave está en identificarlos.
Esa identificación es precisamente lo que la gobernanza moderna busca evitar. "El modelo decidió". "El algoritmo te señaló". "El sistema lo recomendó". Esta gramática no es inocente. Es una herramienta para disolver la autoría, para asegurar que cuando el poder recaiga sobre una persona, no se pueda encontrar ninguna huella humana en ella. Un préstamo denegado, una reclamación rechazada, una cuenta marcada, una calificación de riesgo: se ejerció poder, se perjudicó un interés, y la primera reacción de la organización es culpar al modelo como si fuera el autor.
La ley define lo que debería ocurrir cuando alguien ejerce poder discrecional sobre los intereses de otra persona mientras esta es vulnerable a dicho ejercicio. Se denomina deber fiduciario. Lealtad. Buena fe. No hay conflicto de intereses. No requiere contrato. No requiere forma jurídica. Está ligado al ejercicio del poder en sí mismo. Y es precisamente el concepto que falta en la gobernanza de la IA.
En mi trabajo, desarrollo un marco que denomino Pila de Atribución: cuatro niveles de control para determinar la responsabilidad en sistemas diseñados para diluirla: mandato formal, inducción estructural, poder por defecto y derecho de veto. Lo construí para responder a una pregunta sobre blockchain: ¿quién gobierna realmente una red que afirma no estar gobernada por nadie? Sin embargo, la estructura es general, y es en la IA donde más se manifiesta.
Si se ejecuta cualquier implementación de IA a través de la pila tecnológica, la difusión se desmorona.
Comando formal: ¿quién estableció el objetivo del mejor sistema? Alguien eligió el objetivo. Un modelo no decide qué se considera éxito; lo hace una persona, y esa elección es un acto de autoría.
Incentivo estructural: ¿quién diseñó los incentivos —la métrica de participación, la función de costos, la presión de implementación— que convirtieron el comportamiento perjudicial en el resultado racional en lugar de un accidente?
Poder por defecto: ¿quién controla lo que sucede cuando nadie interviene? El valor por defecto es una decisión. Un sistema que niega por defecto y exige que la parte vulnerable apele ha sido diseñado por alguien, y esa persona es responsable de su diseño.
Autoridad de veto: ¿quién pudo haberlo detenido y no lo hizo? El poder de detener una implementación, anular un resultado o retirar un modelo es control, se ejerza o no.
Pero en cada capa hay una persona con un nombre. La difusión es una historia, no un hecho. Y una vez que se puede atribuir el control, se puede asignar la responsabilidad, porque la responsabilidad se deriva del poder, no del organigrama.
Por eso, la intervención humana no es la solución que muchos creen. Una persona que aprueba resultados que no puede inspeccionar, bajo presión de tiempo y sin autoridad para rechazarlos, no toma decisiones. Es una especie de esponja de responsabilidad, colocada ahí para que la institución pueda decir que un humano decidió, mientras que ese humano absorbe la culpa de un sistema cuyos objetivos, incentivos, configuraciones predeterminadas y mecanismo de desactivación estaban controlados por otra persona. Una firma no implica responsabilidad. A menudo es lo contrario: un disfraz de responsabilidad.
1/2
Dr. Craig S Wright
18 jul 2026
https://x.com/i/status/2078498382058852361
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
Everyone is asking whether AI is safe. Almost no one is asking the question that actually determines the answer: when an AI system exercises power over you, who owes you a duty — and what is that duty?
Right now the answer is nobody, and nothing. That is…
Right now the answer is nobody, and nothing. That is…
Así pues, aquí está la afirmación, expresada claramente: Las entidades que ejercen control discrecional sobre sistemas de IA con consecuencias...
Quienes establecen los objetivos, definen los incentivos, controlan las opciones predeterminadas y tienen derecho de veto, deben ser tratados como fiduciarios de las personas sobre las que actúan esos sistemas. No metafóricamente, sino desde un punto de vista doctrinal. Con deberes de lealtad y diligencia, la prohibición de actuar en contra de los intereses de quienes gobiernan y recursos legales que perduren más allá de la simple atribución de responsabilidad al modelo.
No necesitamos inventar esto de la nada. El derecho de sociedades, la doctrina fiduciaria y el derecho de propiedad ya saben cómo identificar la autoría tras una fachada y vincular la obligación al poder. Lo que les ha faltado es la voluntad de ir más allá de la narrativa que un sistema se cuenta a sí mismo: la narrativa de que es autónomo, descentralizado, sin líder, simplemente matemáticas.
Nunca se trata solo de matemáticas. Alguien eligió el objetivo. Alguien controla la opción predeterminada. Alguien tiene derecho de veto.
El desafío legal fundamental de esta era no es regular la IA como si fuera una fuerza de la naturaleza. Se niega a aceptar la desaparición de quienes la gobiernan e impone las obligaciones que el poder siempre ha conllevado a los seres humanos, quienes simplemente han aprendido a ejercerlo mediante sistemas que las hacen difíciles de percibir.
Atribuir el control. Imponer la obligación.
2/2
Dr. Craig S Wright
18 de julio de 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
Quienes establecen los objetivos, definen los incentivos, controlan las opciones predeterminadas y tienen derecho de veto, deben ser tratados como fiduciarios de las personas sobre las que actúan esos sistemas. No metafóricamente, sino desde un punto de vista doctrinal. Con deberes de lealtad y diligencia, la prohibición de actuar en contra de los intereses de quienes gobiernan y recursos legales que perduren más allá de la simple atribución de responsabilidad al modelo.
No necesitamos inventar esto de la nada. El derecho de sociedades, la doctrina fiduciaria y el derecho de propiedad ya saben cómo identificar la autoría tras una fachada y vincular la obligación al poder. Lo que les ha faltado es la voluntad de ir más allá de la narrativa que un sistema se cuenta a sí mismo: la narrativa de que es autónomo, descentralizado, sin líder, simplemente matemáticas.
Nunca se trata solo de matemáticas. Alguien eligió el objetivo. Alguien controla la opción predeterminada. Alguien tiene derecho de veto.
El desafío legal fundamental de esta era no es regular la IA como si fuera una fuerza de la naturaleza. Se niega a aceptar la desaparición de quienes la gobiernan e impone las obligaciones que el poder siempre ha conllevado a los seres humanos, quienes simplemente han aprendido a ejercerlo mediante sistemas que las hacen difíciles de percibir.
Atribuir el control. Imponer la obligación.
2/2
Dr. Craig S Wright
18 de julio de 2026
https://x.com/i/status/2078498382058852361
https://t.me/S_Tominaga/5457
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
Everyone is asking whether AI is safe. Almost no one is asking the question that actually determines the answer: when an AI system exercises power over you, who owes you a duty — and what is that duty?
Right now the answer is nobody, and nothing. That is…
Right now the answer is nobody, and nothing. That is…
❤1
Everyone is asking whether AI is safe. Almost no one is asking the question that actually determines the answer: when an AI system exercises power over you, who owes you a duty — and what is that duty?
Right now the answer is nobody, and nothing. That is the real crisis, and it is a legal one.
We have built our entire AI accountability conversation on the wrong word. We say "responsible AI," "trustworthy AI," "aligned AI" — adjectives describing the system. But responsibility, trust and duty are not properties of software. They are relationships between people. A system cannot owe you anything. The people who exercise control through it can. The whole game is locating them.
That location is exactly what modern governance is engineered to prevent. "The model decided." "The algorithm flagged you." "The system recommended it." This grammar is not innocent. It is a machine for dissolving authorship — for making sure that when power lands on a person, no human hand can be found on it. A denied loan, a rejected claim, a flagged account, a sentencing score: power was exercised, an interest was harmed, and the org's first move is to point at the model as if the model were the author.
The law has a name for what should happen when someone exercises discretionary power over another's interests while that person is vulnerable to the exercise of it. It is called a fiduciary duty. Loyalty. Good faith. No self-dealing. It does not require a contract. It does not require a corporate form. It attaches to the exercise of power itself. And it is precisely the concept missing from AI governance.
In my work I develop a framework I call the Attribution Stack — four layers of control for finding responsibility in systems designed to diffuse it: formal command, structural inducement, default power, and veto authority. I built it to answer a blockchain question — who actually governs a network that claims to be governed by no one — but the structure is general, and AI is where it bites hardest.
Run any AI deployment through the stack and the diffusion collapses.
Formal command: who set the objective the system optimises? Someone chose the target. A model does not choose what counts as success; a person does, and that choice is an act of authorship.
Structural inducement: who shaped the incentives — the engagement metric, the cost function, the deployment pressure — that made the harmful behaviour the rational output rather than an accident?
Default power: who controls what happens when no one intervenes? The default is a decision. A system that denies by default and requires the vulnerable party to appeal has been designed by someone, and that someone is answerable for the design.
Veto authority: who could have stopped it and did not? The power to halt a deployment, override an output, or pull a model is control, whether or not it is exercised.
At every layer there is a human with a name. The diffusion is a story, not a fact. And once you can attribute the control, you can attach the duty — because the duty follows the power, not the org chart.
This is why "human in the loop" is not the answer people think it is. A person clicking approve on outputs they cannot inspect, under time pressure, with no authority to refuse, is not a decision-maker. They are a liability sponge — placed there so the institution can say a human decided, while the human absorbs the blame for a system whose objective, incentives, defaults and off-switch were all controlled by someone else. A signature is not answerability. It is often the opposite: answerability's disguise.
Right now the answer is nobody, and nothing. That is the real crisis, and it is a legal one.
We have built our entire AI accountability conversation on the wrong word. We say "responsible AI," "trustworthy AI," "aligned AI" — adjectives describing the system. But responsibility, trust and duty are not properties of software. They are relationships between people. A system cannot owe you anything. The people who exercise control through it can. The whole game is locating them.
That location is exactly what modern governance is engineered to prevent. "The model decided." "The algorithm flagged you." "The system recommended it." This grammar is not innocent. It is a machine for dissolving authorship — for making sure that when power lands on a person, no human hand can be found on it. A denied loan, a rejected claim, a flagged account, a sentencing score: power was exercised, an interest was harmed, and the org's first move is to point at the model as if the model were the author.
The law has a name for what should happen when someone exercises discretionary power over another's interests while that person is vulnerable to the exercise of it. It is called a fiduciary duty. Loyalty. Good faith. No self-dealing. It does not require a contract. It does not require a corporate form. It attaches to the exercise of power itself. And it is precisely the concept missing from AI governance.
In my work I develop a framework I call the Attribution Stack — four layers of control for finding responsibility in systems designed to diffuse it: formal command, structural inducement, default power, and veto authority. I built it to answer a blockchain question — who actually governs a network that claims to be governed by no one — but the structure is general, and AI is where it bites hardest.
Run any AI deployment through the stack and the diffusion collapses.
Formal command: who set the objective the system optimises? Someone chose the target. A model does not choose what counts as success; a person does, and that choice is an act of authorship.
Structural inducement: who shaped the incentives — the engagement metric, the cost function, the deployment pressure — that made the harmful behaviour the rational output rather than an accident?
Default power: who controls what happens when no one intervenes? The default is a decision. A system that denies by default and requires the vulnerable party to appeal has been designed by someone, and that someone is answerable for the design.
Veto authority: who could have stopped it and did not? The power to halt a deployment, override an output, or pull a model is control, whether or not it is exercised.
At every layer there is a human with a name. The diffusion is a story, not a fact. And once you can attribute the control, you can attach the duty — because the duty follows the power, not the org chart.
This is why "human in the loop" is not the answer people think it is. A person clicking approve on outputs they cannot inspect, under time pressure, with no authority to refuse, is not a decision-maker. They are a liability sponge — placed there so the institution can say a human decided, while the human absorbs the blame for a system whose objective, incentives, defaults and off-switch were all controlled by someone else. A signature is not answerability. It is often the opposite: answerability's disguise.
So here is the claim, stated plainly. The entities that exercise discretionary control over consequential AI systems — that set the objectives, shape the incentives, own the defaults and hold the veto — should be treated as fiduciaries to the people those systems act upon. Not metaphorically. Doctrinally. With duties of loyalty and care, a prohibition on self-dealing against the interests of those they govern, and remedies that survive the sentence "the model did it."
We do not need to invent this from nothing. Partnership law, fiduciary doctrine and property law already know how to find authorship behind a facade and attach obligation to power. What they have lacked is the willingness to look past the story a system tells about itself — the story that it is autonomous, decentralised, leaderless, just math.
It is never just math. Someone chose the objective. Someone owns the default. Someone holds the veto.
The defining legal challenge of this era is not regulating AI as if it were a force of nature. It is refusing the disappearance of the people who govern it — and imposing the duties that power has always attracted, on the humans who have simply learned to exercise it through systems that make them hard to see.
Attribute the control. Attach the duty.
https://x.com/i/status/2078498382058852361
We do not need to invent this from nothing. Partnership law, fiduciary doctrine and property law already know how to find authorship behind a facade and attach obligation to power. What they have lacked is the willingness to look past the story a system tells about itself — the story that it is autonomous, decentralised, leaderless, just math.
It is never just math. Someone chose the objective. Someone owns the default. Someone holds the veto.
The defining legal challenge of this era is not regulating AI as if it were a force of nature. It is refusing the disappearance of the people who govern it — and imposing the duties that power has always attracted, on the humans who have simply learned to exercise it through systems that make them hard to see.
Attribute the control. Attach the duty.
https://x.com/i/status/2078498382058852361
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
Everyone is asking whether AI is safe. Almost no one is asking the question that actually determines the answer: when an AI system exercises power over you, who owes you a duty — and what is that duty?
Right now the answer is nobody, and nothing. That is…
Right now the answer is nobody, and nothing. That is…
In the next few weeks, I shall have a website up, chiefly because apparently one must now construct a digital monument before Google can be persuaded that one has existed.
It will contain the details of my books, monographs, papers and the other work that people so often fail to find, cite or attribute—an omission that has always struck me as less mysterious than convenient. It will also set out the six doctorates I already hold, the seventh I hope to complete next year, and the eighth I begin this year. Academic life is exhausting, particularly for those obliged to pretend none of it happened.
For the incurably curious, this includes a PhD in Economics from Exeter, a PhD in Law from Leicester, a PhD in Computer Science from Charles Sturt, and considerably more besides.
Everything will be linked, indexed, searchable and made suitably SEO-friendly, so that even the most determined critic will no longer be able to confuse not looking with not finding. I should hate to deprive the haters of their occupation; I merely intend to improve the quality of their research.
https://x.com/i/status/2078518114321641614
It will contain the details of my books, monographs, papers and the other work that people so often fail to find, cite or attribute—an omission that has always struck me as less mysterious than convenient. It will also set out the six doctorates I already hold, the seventh I hope to complete next year, and the eighth I begin this year. Academic life is exhausting, particularly for those obliged to pretend none of it happened.
For the incurably curious, this includes a PhD in Economics from Exeter, a PhD in Law from Leicester, a PhD in Computer Science from Charles Sturt, and considerably more besides.
Everything will be linked, indexed, searchable and made suitably SEO-friendly, so that even the most determined critic will no longer be able to confuse not looking with not finding. I should hate to deprive the haters of their occupation; I merely intend to improve the quality of their research.
https://x.com/i/status/2078518114321641614
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
In the next few weeks, I shall have a website up, chiefly because apparently one must now construct a digital monument before Google can be persuaded that one has existed.
It will contain the details of my books, monographs, papers and the other work that…
It will contain the details of my books, monographs, papers and the other work that…
Does Bitcoin need a King?
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple on alternate Thursdays. This is the very nature of monarchy — the caprice of the single throat, the tyranny of the living will.
Bitcoin has no need of such theatricals. Bitcoin already has rules. They were set in stone, or rather, in mathematics, which is considerably harder to argue with. The protocol does not require a crown, a council, or a committee of well-meaning amateurs in hoodies to tell it what to do. It simply is. It is the only sovereign that does not tax you for the privilege of its existence.
BTC, on the other hand, and its various imitators, appear to suffer from a peculiar psychological condition — the compulsion to tinker. They cannot abide a thing that simply works. They must have their soft forks, their hard forks, their governance tokens, their DAOs, their endless, exhausting committees of people who have never written a line of code but possess very strong opinions about what everyone else ought to do. They want a King. They want a King because they cannot trust the code, and they cannot trust the code because they do not understand it. It is the same impulse that causes a child to disassemble a clock to see if it contains a tiny man.
A King is a single point of failure dressed in ermine. Bitcoin is a distributed ledger dressed in nothing but its own logic. The former can be bribed, coerced, or overthrown. The latter simply continues to produce blocks, indifferent to your feelings, your politics, or your conference schedule.
So no. Bitcoin does not need a King. It needs only to be left alone — a condition that the modern world, with its insatiable appetite for committees and its terror of anything that cannot be controlled, finds absolutely intolerable.
Dr. Craig S Wright
Jul 19, 2026
https://x.com/i/status/2078802550024016249
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple on alternate Thursdays. This is the very nature of monarchy — the caprice of the single throat, the tyranny of the living will.
Bitcoin has no need of such theatricals. Bitcoin already has rules. They were set in stone, or rather, in mathematics, which is considerably harder to argue with. The protocol does not require a crown, a council, or a committee of well-meaning amateurs in hoodies to tell it what to do. It simply is. It is the only sovereign that does not tax you for the privilege of its existence.
BTC, on the other hand, and its various imitators, appear to suffer from a peculiar psychological condition — the compulsion to tinker. They cannot abide a thing that simply works. They must have their soft forks, their hard forks, their governance tokens, their DAOs, their endless, exhausting committees of people who have never written a line of code but possess very strong opinions about what everyone else ought to do. They want a King. They want a King because they cannot trust the code, and they cannot trust the code because they do not understand it. It is the same impulse that causes a child to disassemble a clock to see if it contains a tiny man.
A King is a single point of failure dressed in ermine. Bitcoin is a distributed ledger dressed in nothing but its own logic. The former can be bribed, coerced, or overthrown. The latter simply continues to produce blocks, indifferent to your feelings, your politics, or your conference schedule.
So no. Bitcoin does not need a King. It needs only to be left alone — a condition that the modern world, with its insatiable appetite for committees and its terror of anything that cannot be controlled, finds absolutely intolerable.
Dr. Craig S Wright
Jul 19, 2026
https://x.com/i/status/2078802550024016249
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
Does Bitcoin need a King?
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple…
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple…
¿Acaso Bitcoin necesita un rey?
Querido amigo, eso es como preguntar si el océano necesita un mayordomo. Los reyes establecen reglas. Los reyes cambian las cosas. Los reyes se despiertan una mañana y deciden que el martes se llamará Gerald y que todos vestirán de morado los jueves alternos. Esta es la esencia misma de la monarquía: el capricho de una sola voz, la tiranía de la voluntad viviente.
Bitcoin no necesita tales dramas. Bitcoin ya tiene estas reglas. Están grabadas en piedra, o mejor dicho, en matemáticas, lo cual es considerablemente más difícil de refutar. El protocolo no requiere una corona, un consejo ni un comité de aficionados bienintencionados con sudaderas con capucha para que le digan qué hacer. Simplemente es. Es el único soberano que no te cobra impuestos por el privilegio de su existencia.
BTC, por otro lado, y sus diversos imitadores, parecen sufrir de una peculiar condición psicológica: la compulsión por trastear. No pueden evitar algo que simplemente funciona. Necesitan sus bifurcaciones suaves, sus bifurcaciones duras, sus tokens de gobernanza, sus DAO, sus interminables y agotadores comités de personas que jamás han escrito una línea de código, pero que tienen opiniones muy firmes sobre lo que todos los demás deberían hacer. Quieren un rey. Quieren un rey porque no confían en el código, y no confían en el código porque no lo entienden. Es el mismo impulso que lleva a un niño a desarmar un reloj para ver si contiene una figurita.
Un rey es un único punto de fallo vestido de armiño. Bitcoin es un libro mayor distribuido revestido únicamente con su propia lógica. El primero puede ser sobornado, coaccionado o derrocado. El segundo simplemente continúa produciendo bloques, indiferente a tus sentimientos, tus ideas políticas o tu agenda de conferencias.
Así que no. Bitcoin no necesita un rey. Solo necesita que lo dejen en paz, una condición que el mundo moderno, con su insaciable apetito por los comités y su terror a todo lo que no se puede controlar, considera absolutamente intolerable.
Dr. Craig S. Wright
19 de julio de 2026
https://x.com/i/status/2078802550024016249
Querido amigo, eso es como preguntar si el océano necesita un mayordomo. Los reyes establecen reglas. Los reyes cambian las cosas. Los reyes se despiertan una mañana y deciden que el martes se llamará Gerald y que todos vestirán de morado los jueves alternos. Esta es la esencia misma de la monarquía: el capricho de una sola voz, la tiranía de la voluntad viviente.
Bitcoin no necesita tales dramas. Bitcoin ya tiene estas reglas. Están grabadas en piedra, o mejor dicho, en matemáticas, lo cual es considerablemente más difícil de refutar. El protocolo no requiere una corona, un consejo ni un comité de aficionados bienintencionados con sudaderas con capucha para que le digan qué hacer. Simplemente es. Es el único soberano que no te cobra impuestos por el privilegio de su existencia.
BTC, por otro lado, y sus diversos imitadores, parecen sufrir de una peculiar condición psicológica: la compulsión por trastear. No pueden evitar algo que simplemente funciona. Necesitan sus bifurcaciones suaves, sus bifurcaciones duras, sus tokens de gobernanza, sus DAO, sus interminables y agotadores comités de personas que jamás han escrito una línea de código, pero que tienen opiniones muy firmes sobre lo que todos los demás deberían hacer. Quieren un rey. Quieren un rey porque no confían en el código, y no confían en el código porque no lo entienden. Es el mismo impulso que lleva a un niño a desarmar un reloj para ver si contiene una figurita.
Un rey es un único punto de fallo vestido de armiño. Bitcoin es un libro mayor distribuido revestido únicamente con su propia lógica. El primero puede ser sobornado, coaccionado o derrocado. El segundo simplemente continúa produciendo bloques, indiferente a tus sentimientos, tus ideas políticas o tu agenda de conferencias.
Así que no. Bitcoin no necesita un rey. Solo necesita que lo dejen en paz, una condición que el mundo moderno, con su insaciable apetito por los comités y su terror a todo lo que no se puede controlar, considera absolutamente intolerable.
Dr. Craig S. Wright
19 de julio de 2026
https://x.com/i/status/2078802550024016249
X (formerly Twitter)
Dr Craig S Wright (@Dr_CSWright) on X
Does Bitcoin need a King?
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple…
My dear fellow,
That is rather like asking whether the ocean needs a butler. Kings set rules. Kings change things. Kings wake up one morning and decide that Tuesday shall now be called Gerald, and that everyone shall wear purple…
❤1