The argument behind
a commons-centric
architecture.
TLT is built on a specific set of ideas. This is where we lay them out fully: the evidence, the intellectual case, and the structure behind every design decision.
What a shareholder-centric architecture produces for people
The food and health system fails people and the environment consistently and predictably, while consistently delivering for shareholders. That is not bad luck. It is the architecture.
Ten companies account for the vast majority of packaged food brands available in global retail (Oxfam, Behind the Brands, 2013). Their incentive is to sell more, not to nourish better. The healthcare system that manages the consequences follows the same logic: the United States spends more per person on healthcare than any country on earth, and ranks below 40 other nations in life expectancy. That is not underperformance. It is the predictable output of a system optimised for revenue from treatment rather than for the prevention that would make treatment unnecessary.
The digital health and food assistants and trackers that emerged to address these failures did not change the underlying logic. They replicated it. A platform whose revenue depends on engagement has a structural incentive to keep users returning, not to make them healthy enough to need it less. Every major platform in health and food is caught in the same model: accumulate users, extract value from their data, and grow until they own the category. The medium shifted from shelf space and hospital contracts to algorithms and notification psychology. The incentive structure did not.
One might expect this to moderate as organisations grow large enough and their founders wealthy enough. It does not self-correct. Wang, Jetten and Steffens tracked more than 175,000 people across 78 countries and found that wealth does not reduce the drive to accumulate. It intensifies it. As platforms grow, the people who run them tie their psychological identity increasingly to what they own. Dacher Keltner's two decades of laboratory research on power add the neurological detail: holding power measurably reduces the activation of mirror neurons — the biological basis of empathy — while increasing self-referential focus. Power changes what people want, not only what they can do. The pressure to reform the model has to come from outside or not at all.
C. Northcote Parkinson observed the organisational consequence in 1958: institutions expand their complexity and headcount to consume whatever resources they accumulate, independent of actual output. But this is more than cultural drift. It becomes structurally anchored. Departments exist to justify their own budgets. Complexity is added to manage complexity. The organisation's capacity to change direction decreases as its size increases. What began as a business decision hardens into an institutional constraint that no individual inside the structure has the leverage to reverse.
Daron Acemoglu's 2024 Nobel work, formalised in Power and Progress (2023, with Simon Johnson), establishes the economic endpoint: monopoly power, once entrenched, first reduces innovation, then redirects what remains. It moves away from prevention, away from broadly beneficial public health, toward marginally improved, patent-protected products with maximum pricing power. The extraction deepens. The monetisation narrows. The majority of people, who need prevention and affordability most, receive progressively less of both.
Why a commons-centric alternative is structurally superior
The failure is structural. So is the solution, but only when the human is truly at the centre. Not as a product. As a person.
Why distributed governance, constrained optimisation, and preserved adaptive capacity produce more value than any centralised architecture can.
Why the two obvious alternatives, state regulation and market competition, both fail for structural rather than incidental reasons.
How TLT applies the commons architecture in practice, and why every design decision follows from the argument above.
Our position on AI: Whose AI?
The most discussed and consequential technology of our time. Compared to the industrial revolution: 10× as fast (10y) and 10× as much impact. Among the least governed.
The dominant AI architecture (LLM) extracts data from many, concentrates intelligence in few. It bets the entire value of the technology on being first. First to what? Exactly! Accountable to whom? Exactly! These are governance questions. The people moving fastest are not stopping to ask them.
The shoulders
we're standing on.
TLT did not invent any of these ideas. But we are applying them to food, to data, to how a small organisation might behave differently if it builds the right constraints into its own architecture from the start.
These thinkers did not know each other's work in many cases, and none of them set out to design a food-tracking app. What is striking is how independently they arrived at structurally similar conclusions: that viable, productive systems require distributed governance, constrained optimisation, and the preservation of adaptive capacity. We are applying a convergent set of insights from biology, cybernetics, economics, political science, and psychology to a specific problem in digital health.
If this resonates, the app is the first step.
Everything described here is built into the architecture. The app is how you begin contributing to it.
