"Gonna Start a Revolution From My Bed"

With apologies and deep respect to the Gallaghers · Don't Look Back in Anger, Oasis, 1995

midagent AI · April 2026 John Pope · Founder, midagent AI · Ottawa, Canada

https://youtu.be/MpNHqJs7rXQ?si=-LxACXf8wNT1P9eP

The wisdom of positive sum expressed in melody and verse.

I want to be honest with you about something.

When I decided to start this company, I was not sitting in a boardroom. I was not emerging from a successful exit with capital to deploy and a network of co-investors on speed dial. I was doing what a lot of people who have become furious about something do: I was thinking, reading, and getting increasingly agitated in the specific way that happens when you see a problem clearly enough to know it is solvable and yet watch it go unsolved year after year because nobody with sufficient standing has decided to care.

So yes. In a fairly literal sense, I started a revolution from my bed.

Noel Gallagher wrote that lyric in 1995 as a provocation — the absurdist self-awareness of the dreamer who knows the dream is audacious and says it anyway. I have always loved it for the same reason I love all honest admissions of improbability: they are more credible than confident declarations, and credibility is what every revolution actually runs on.

This one is non-violent. Obviously. The weapons are architecture, governance, and economic logic. The territory being contested is not physical. It is the infrastructure layer of the global digital economy — the platforms through which commerce flows, the systems through which governments process their most sensitive intelligence, and the data pipelines through which the raw material of the AI age is collected, refined, and monetized. And the revolution, if I am being precise about it, is not against any particular company or country. It is against a set of structural arrangements that have become, quietly and without anyone quite noticing, deeply incompatible with the world we need to build next.

The Trust Recession Nobody Named

Here is the paradox at the centre of this moment in history. We are about to deploy the most powerful decision-making technology ever created — artificial intelligence systems that will diagnose disease, allocate capital, drive vehicles, manage power grids, coordinate supply chains, and assist the analytical work of governments across every domain of public policy. And we are doing it at the precise historical moment when trust in the institutions meant to govern that technology is at a generational low.

Trust in American leadership — the bedrock assumption of the post-war international order — is eroding with a speed that would have been unimaginable five years ago. The governments of Canada, Germany, Japan, Australia, and the United Kingdom are not merely recalibrating their trade relationships. They are fundamentally reconsidering whether the digital infrastructure they have built on American corporate foundations is compatible with their own sovereignty, their own legal systems, and their own national interests. These are not fringe concerns. They are mainstream policy preoccupations in every G-Middle capital.

At the same time, a broader movement is gathering — not organized by any single institution, not led by any single figure, but unmistakably present in the conversations happening in parliaments, in civil society, in research institutes, and among the technologists themselves who are increasingly vocal about what they have built and what it might become. It is a movement of people who are genuinely anxious about the direction of AI development, who believe that the concentration of AI capability in a small number of unaccountable corporations in a single jurisdiction is a structural problem rather than a temporary inconvenience, and who are demanding something that political systems have historically struggled to provide on the required timeline: responsible governance of a technology that is moving faster than the governance frameworks designed to contain it.

The EU AI Act. The UK's AI Safety Institute. Canada's Voluntary Code of Conduct for Generative AI. The G7 Hiroshima AI Process. The UN's AI Advisory Body. Geoffrey Hinton — the Godfather of AI, trained at the University of Toronto — leaving Google to speak freely about existential risk. These are not the signs of a technology being welcomed with open arms. They are the signs of a civilization trying urgently to get ahead of something it is not sure it can control.

What is missing from almost every one of these governance efforts is the thing that governance frameworks most require to function: trustworthy infrastructure to govern. You cannot regulate a system you cannot see. You cannot protect citizens from an architecture whose design is proprietary, whose decision logic is opaque, and whose legal jurisdiction is foreign. The governance ambition is real. The infrastructure to make it possible does not yet exist at scale. That is the gap this revolution is trying to fill.

Why e-Commerce, Advertising, and Data Analytics?

This is a fair question and it deserves a direct answer. If the concern is AI governance and the future of human civilization, why is the practical focus of this project on merchant fees, digital advertising markets, and government analytics contracts? The answer has two parts, one strategic and one moral.

The strategic answer is that the commerce layer is the beachhead. The infrastructure through which AI agents will coordinate the global economy — purchasing decisions, supply chain management, price discovery, inventory optimization, demand forecasting — is being built right now, in the decisions being made today about which platforms handle which transactions on what terms. The platform that owns the commercial interface of the agentic economy owns the most consequential data pipeline in human history: a real-time record of what every business produces, what every consumer wants, what every supply chain can deliver, and how every market clears. That is not a commerce dataset. That is the training corpus for the economic intelligence layer of artificial general intelligence. Whoever owns the commerce rail of the agentic economy owns something considerably more significant than a marketplace.

The moral answer is simpler. The people being hurt by the current arrangement are not abstractions. They are the 1.19 million Canadian SME owners who are watching their margins consumed by platform fees they cannot negotiate and cannot escape. They are the restaurant operators working 70-hour weeks to produce $1.80 of profit from a $60 delivery order. They are the government employees whose sensitive analytical work is being processed on infrastructure subject to foreign legal compulsion. They are, in Carney's phrase, the people trying to afford a Canada that is becoming increasingly unaffordable partly because the infrastructure of the digital economy is designed to extract value from them rather than create value with them.

The revolution starts in e-commerce and digital advertising because that is where the extraction is most measurable, most immediate, and most correctable. But the revolution does not end there.

What a Trust Architecture Actually Is

I want to introduce you to a phrase that I think is going to become one of the defining concepts of the AI age: trust architecture.

We are accustomed to thinking about trust as a property of relationships — something that builds over time between people and institutions through consistent behaviour, honesty, and demonstrated alignment of interests. That kind of trust is real and important. But it has a critical vulnerability: it depends on the ongoing character of the parties involved. Change the leadership, change the incentive structure, face a sufficiently large acquisition offer or a sufficiently acute commercial pressure, and claimed trust meets its test.

A trust architecture is different. It is trust embedded in the design of the system itself — in the governance instruments, the technical standards, the legal structures, and the commercial models that make trustworthy behaviour not a matter of ongoing choice but a structural property of how the system operates. You cannot build a trustworthy AI economy by asking powerful institutions to choose to be trustworthy. You build it by designing systems in which trustworthy behaviour is the only available option, in which the mechanisms for exploitation have been architecturally disabled, and in which every stakeholder can verify the trustworthiness of the system independently without relying on the representations of the party running it.

This is what Project Sovereign Nexus has been designed to be. Not a company that claims to be trustworthy, but a system whose trustworthiness is verifiable at every layer — governance, technical, legal, commercial, and human. We have documented this architecture in detail, for anyone who wants to examine it, in a white paper called The Architecture of Trust. It is available at midagent.ca, and it is written to be challenged, not accepted on faith. That, in itself, is the point.

The Movement Is Already Here

More than 30 countries have enacted or are actively developing AI governance frameworks. The EU AI Act — the world's first comprehensive AI regulation — is in force. The UK, Canada, Japan, and Australia are all pursuing bilateral AI safety cooperation. Civil society organizations across the G-Middle are publishing frameworks for algorithmic accountability, data sovereignty, and responsible AI deployment.

The open-source AI community has produced Meta's Llama, Mistral and many others. Cohere's Command R+ serves Canada and other nations as a credible sovereign alternative to closed proprietary models from the US. The Model Context Protocol has achieved universal major-provider adoption as an open standard in sixteen months. Academic AI safety research has expanded from a niche sub-field to a mainstream preoccupation with dedicated institutes at Oxford, Cambridge, MIT, and UC Berkeley.

This is not a fringe movement flailing its radical hands in the air. It is the rational, mainstream response of a civilization that is beginning to take seriously the stakes of getting AI wrong. And acting responsibly to mitigate against those downside risks.

Positive-Sum or Nothing

The fundamental choice confronting every institution, every government, and every technologist working on AI right now is between two value systems that produce radically different outcomes.

The zero-sum approach treats AI as a competitive weapon: the nation or corporation that achieves dominance first wins, and winning means extracting maximum value from every other participant in the system. This approach is not irrational, given the incentive structures that currently govern both corporate strategy and great-power competition. It is, however, genuinely dangerous — because a zero-sum AI race between major powers, conducted on infrastructure owned by the combatants and governed by frameworks designed to serve their interests, produces exactly the kind of arms race dynamic that makes the worst outcomes more likely.

The positive-sum approach treats AI as shared infrastructure whose value is maximized when it is deployed in systems that create value for all participants rather than extracting it from some for the benefit of others. This is not naïve idealism. It is the same logic that produced the internet, the World Wide Web, the open-source software movement, and every other general-purpose technology whose transformative impact derived precisely from its open, interoperable, non-extractive architecture. The technologies that changed everything were not the ones that maximized rent extraction from their users. They were the ones that maximized the value available to their users — and captured a thin, sustainable return on the infrastructure that made it possible.

Utility pricing for digital commerce is positive-sum. Open standards for AI agent interoperability are positive-sum. Data sovereignty frameworks that return the economic value of personal data to the individuals who created it are positive-sum. Governance institutions that are designed to serve the public interest rather than the interests of the platforms they govern are positive-sum. Trust architectures that make exploitation structurally impossible rather than merely discouraged are positive-sum.

These are not abstract values. They are specific design choices that can be made or not made, right now, in the infrastructure decisions that will shape the AI economy for the next fifty years. The window for making them — before the extractive architectures are so deeply embedded that displacing them becomes prohibitively expensive — is open. It will not remain open indefinitely.

The Stakes of Getting This Wrong

I do not think it is alarmist to say that the decisions being made in the next five years about how AI infrastructure is owned, governed, and deployed will be among the most consequential decisions in human history. I think it is accurate.

The generation that is alive today is the first generation to face this specific choice: to build the AI age on a foundation of concentrated, extractive, opacity-dependent infrastructure — or to build it on a foundation of distributed, trustworthy, transparent, positive-sum architecture. Every previous generation faced consequential choices. None of them faced this one, because none of them had this technology.

The environmental movement has a phrase for the moral weight of this kind of decision: intergenerational equity. The idea that the choices made by the current generation impose consequences on generations that have no voice in making them, and that this asymmetry creates an obligation to think beyond immediate self-interest.

The fossil fuel economy created an intergenerational equity problem in the domain of climate. The AI economy has the potential to create one in the domain of power: a world in which a small number of entities, having established dominance over the intelligence infrastructure of the global economy during the critical window of the early AI age, exercise a form of structural power over subsequent generations that those generations will find very difficult to dismantle.

That is the future we are trying to prevent. Not through protest, not through regulation alone, but through the patient, rigorous work of building trustworthy alternatives before the extractive ones become too entrenched to displace.

I started this company from a place of genuine alarm about the direction of things, and a genuine conviction that the direction could be changed — not by asking powerful institutions to behave differently, but by building institutions whose architecture makes the behaviour we need the only available option.

I am not the only one who believes this. The movement of people, companies, civil society organizations, researchers, and governments demanding digital sovereignty, responsible AI governance, and trustworthy infrastructure is real, it is growing, and it is beginning to produce tangible results: regulations with teeth, open standards with adoption, sovereign infrastructure with funding, and governance frameworks with genuine accountability.

The revolution is non-violent and it is already underway. It is being conducted in architecture documents and trust deed provisions and open-source repositories and bilateral diplomatic agreements and the quiet decision of a million merchants to ask whether the platform they depend on is actually on their side.

We are building for the generation that comes after us — the one that will inherit whatever architecture we embed in the AI age, for better or for worse. We intend to give them something worth inheriting: an economy that serves everyone, infrastructure that encodes trust rather than extraction, and a governance model that proves positive-sum values are not just admirable but durable.

So slip inside the eye of your mind — and don't look back in anger. Not because the present is perfect. But because the future is still ours to build.