“Even if everything that traditional futurists have asked for is delivered — eternal youth, unlimited material wealth, morphological freedom, superintelligence, immersive VR, molecular nanotechnology, etc. — there is no evidence that our subjective quality of life would on average significantly surpass the quality of life of […] a New Guinea tribesman today.”

— David Pearce, The Abolitionist Project

Almost every serious conversation about the future — AI, climate, geopolitics, longevity — is a conversation about rearranging the world outside the skull. Very little of it concerns the underlying brain-mechanism that decides how any arrangement will actually feel from the inside.

There is a vision aimed at these. It is very un-prominent and doesn’t even have a standard name. In this post I’ll call it Paradise Engineering.

Before dismissing this vision, it is worth looking at the underlying first principles of happiness.

One: humans chase positive states. Earned. Delicious. Fascinating. Beautiful. Satisfying. We have an enormous, finely graded lexicon for how things feel to us, and a comparatively crude one for what things are. The language tells you what we are actually tracking.

This doesn’t mean that we can’t act upon principles and altruistically, but positive emotions are the main driver of our conscious action. Just think of eating, consuming, creating or your own dream-life.

Two: those states are neural events. Whether something is earned or delicious or fascinating is not a property of the thing. It is a description of what your brain did with sensory and contextual input.

Consider “attractive.” Show a baboon a photograph of a woman generally regarded as beautiful and he will find her unremarkable at best. The feeling is mutual. Nothing about attractiveness lives in the object; it lives in a specific evolved circuit, and other circuits run other software. Or take depression, where a life with no external problem at all can be experienced as unbearable. The world stays the same. The processing changes, and with it everything that matters.

Three: neural events are physical, and physics is editable. Thought and perception run on action potentials, which run on ions moving across membranes. The brain mechanics are very complicated, yes. But not magic. And the tools to manipulate brain mechanisms are rapidly evolving: SSRIs, ketamine, transcranial stimulation, deep brain stimulation, gene editing, brain–computer interfaces.

The conclusion: if how good your life feels is a physical state, it is a state that can, in principle, be set.

“The world of the superhappy is quite different from the human world.”

Billions of animals are killed every year to produce a few seconds of pleasant activity in a human mouth and a human reward circuit. Billionaires buy superyachts and private jets for a status signal that fades within weeks. Children scroll until two in the morning for a series of small spikes in the ventral striatum. In each case we are paying an enormous external cost — ecological, moral, personal — for a tiny, temporary internal gain.

Read that way, paradise engineering is not a drug fantasy. It is an efficiency argument. We are already spending the world on our reward circuits. We are simply doing it with catastrophic overhead, because the only lever we ever had was the world.

Nobody needs to shovel sugar or drink themselves numb to feel all right. Unconditional wellbeing paired with intact concern for ecological, social and intergenerational problems is not obviously a worse configuration than the one we have. We might create a society that is more happy, healthy and responsible.

Wireheading

“A 48-year-old woman with a stimulating electrode implanted in the right thalamic nucleus ventralis posterolateralis developed compulsive self-stimulation associated with erotic sensations and changes in autonomic and neurologic function. This case establishes the potential for addiction to deep brain stimulation […].”

This is an excerpt from the abstract of a study from 1986. The patient ended up self-stimulating throughout the day, “neglecting personal hygiene and family commitments,” through cardiac episodes and neurological symptoms. By the time of the study she had lost eleven IQ points.

David Pearce, the philosopher behind the term “paradise engineering,” knows this and rejects it. His proposal is not a permanent orgasm; it is “information-sensitive gradients of bliss.” You would still prefer some things to others, still be moved, still be motivated. Everything would simply happen above zero. Whether that is engineerable is an open question!

Are emotions the destination, or the instrument we navigate by?

The suspicion in regard to paradise engineering is that suffering carries information we cannot afford to lose: about ourselves, about each other, about what makes a moment worth living.

Think for example of loneliness. On the one hand loneliness is a huge problem and a really bad feeling. On the other hand it moves people to socialise, creating connections and shared experiences.

Changing the map and compass means changing purpose and therefore humanity. That is not necessarily wrong: humanity is not the crown of creation.

But if we start altering our navigation system for a meaningful life, we need an adequate answer to how it gets recalibrated. If the compass itself is re-aimed, which seems likely, we need an adequate answer to where it should point.

For all of human history, the question “how do we feel better?” had exactly one available answer: rearrange the world. That constraint was so total that we mistook it for the question itself. It is now dissolving, and we may need to answer the question of meaning in the first place.

• In the first part, I will outline the structural problems of representative democracy.
• In the second part, I will derive requirements for a better state from these problems.
• In the third part, I will construct a new concept of governance that should lay the foundation for the politics of the 21st century.

Part One: Structural Problems of Democracies

• Political disenchantment
• Cognitive biases and limitations
• Exclusivity

Representative politics is in a deep crisis. The most powerful man in the world, Trump, is a populist who emotionalises people and drives division. His policies are debatable, but they are by no means optimal and there are obvious failures (e.g. the Iran war). On the world stage, we can see how representative politics fails and looks away when faced with long-term crises such as climate change.

These problems are not distant from us or immutable — they are a product of the voting decisions made by you and me. The problem lies within the system of (representative) democracy.

No Interest and No Knowledge — Political Disenchantment

The majority of voters have little interest in politics and little knowledge of it. This leads to poor and unqualified voting decisions. Nuanced and unpopular arguments and decisions that are “correct” cannot and will not be understood by these people, and therefore fail to find majorities. The foundation of democracy — participation and engagement — is eroding. Populists and simple explanations fill the void.

No Rational Decisions

Voters and politicians alike are human beings, not machines. This leads to limitations and distortions:

1. Information: Voters and politicians can only process a very limited amount of information. The majority of voters consequently make decisions on an extremely narrow information base. This already very limited information base is shaped by the media, which is itself subject to distortions and frequently serves a particular clientele. Because media companies pursue commercial interests and the viewer is simultaneously always a consumer, the cognitive biases of people are further amplified (emotionalisation, subjectivity over objectivity, confirmation of existing opinions). In its most extreme form, this manifests in social media algorithms that reinforce opinions and interests within echo chambers. This influence through the niche-ification of discourse leads to strongly correlated opinion formation and renders the principle of the “Wisdom of the Crowds” — often invoked in defence of democratic systems — effectively inoperative.
2. Information processing: Even if the quality of information were perfect and people had unlimited capacity to process it, maximally rational decisions would still be impossible due to the cognitive biases of voters and politicians. At the core of this finding is the recognition that human beings are simply not computers. They are shaped by evolution, which is why they prioritise short-term interests over long-term ones. At its core, the processing of information does not follow the objective, mathematical trade-offs that good policy would require, but rather subjectively weighted arguments and opinions.

For an overview of cognitive biases that significantly impair rational thinking: en.wikipedia.org/wiki/List_of_cognitive_biases

An example — the “Present Bias”:

People are subject to what is known as the “Present Bias.” This describes a distortion in which people prioritise short-term interests over long-term ones. Most people will recognise this from everyday life: it takes discipline and awareness not to eat unhealthy food (such as a pizza). It provides immediate gratification, but in the long run leads to obesity and increased risk of disease.

In politics, this manifests as “short-termism”:

• Cheap conventional energy is prioritised over preserving balanced climate systems
• Short-term consumer debt is prioritised over maintaining fiscally healthy and equitable public finances in the long run
• Short-term stabilisation of pension, healthcare, and infrastructure systems is prioritised over far-sighted restructuring and investment

When I confronted a senior politician with the obvious need to reform Germany’s statutory pension system (raising the retirement age), he agreed with me on the substance, but said it was something “you can’t win elections with.” This illustrates that far-sighted policy in democracies is structurally sacrificed to short-term and convenient interests.

These biases and limitations are not a minor problem that can be resolved through reflection — they are structurally embedded in the media system, in human cognitive capacity, and in the very way we think. Democratic systems are not well suited to objective, rational, and unpopular decision-making.

Presentocracy and Anthropocracy

The original democracy of ancient Greece is, in retrospect, highly exclusive and therefore undemocratic. The right to vote was granted only to free men over the age of 20 with an established reputation. As a result, the proportion of eligible voters relative to the total population was approximately 1 in 10. The abolition of slavery, the equal rights of Black people, and the emancipation of women gradually remedied these deficits. Emancipation and equal rights are rooted in a transformed conception of humanity.

For the 21st century, we need a dramatic reform of our worldview. Climate change and the general tendency towards short-term politics demand that the interests of future generations be given a voice within the political system. Animals, too, should be given a voice. This is a logical application of our own moral principles and long overdue.

Part Two: Requirements for a Better State

Inclusion

The political system should no longer serve exclusively the interests of the human group, but should allow future generations and animals to have influence over political decisions. This inclusion is necessary to secure a liveable future and to achieve a broader form of equal representation.

Analysis

Politics should no longer be based on subjective sentiment, which — as described above — is in large part unqualified, distorted, and severely limited. Instead, we need a mechanism that collects available information and analyses it efficiently. This would enable unpopular but correct decisions to be made.

Acceptance

A system that also makes unpopular and potentially radical decisions — ones that may limit the comfort and privileges of all people or of specific groups — must nonetheless take existing power dynamics into account and secure acceptance. Otherwise, even the best and most correct policy amounts to nothing.

The trade-off is that short-term acceptance and satisfaction decline, in exchange for a long-term, more far-sighted, and more inclusive politics being implemented. Current politics concerns itself with the short-term dissatisfactions of citizens and tries to please them. This is not only an accurate description of the style of governance in many countries, but a direct consequence of the “Present Bias.”

To establish such a political system and secure organic acceptance, appropriate principles of participation, transparency, and shared belief must be developed.

Restructuring the Right to Vote

1. Rights of future generations: To make life on Earth possible for future generations, the interests of those generations (e.g. an ecological basis for life) must be given a voice within the political system. Since the interests of these generations currently hold no power or legitimacy, a reform of the political system is required to create that power and legitimacy.
2. Animal rights: In the spirit of our own moral principles (no violence, the right to life, dignity …) it is only logical to include the interests of thinking and feeling animals in politics. The legitimacy to enforce exclusively human interests derives from the power to do so. Animals do not possess that power, which is why either: a) formal and practical power must be granted to the interests of animals, or b) the interests of animals must acquire formal power and social norm through a changed worldview. On closer inspection, it becomes clear that this is a chicken-and-egg problem, and that a shift in worldview will likely only come about through a dramatic shift in power. The equal standing of animals will, in the long run, represent an enormous step forward — comparable to previous emancipation and equal rights reforms.

Part Three: The World Model

To fulfil the requirements outlined above, a world model is a suitable form of political system. This new type of AI model consists of models capable of simulating the world in specific scenarios, at varying levels of abstraction. To do this, they learn physical, social, and organisational regularities from data and construct a model of the world. This world model can then be used to simulate states and actions. Through a cost function, the consequences of a given policy can be simulated in this way (“A Path Towards Autonomous Machine Intelligence”, Yann LeCun, 2022). Even if this is imperfect — the world model is not the world itself, and the technology is still in its infancy — the potential can already be glimpsed.

A world model can in this way turn the best policy within the constraints of interests and costs into an optimisation problem and solve it. Guardrails can be built into the world model to set moral principles and competencies (as a constitutional equivalent). Different scenarios can be transparently weighed against one another.

The only limits of this world model — and of policy quality — are the limits of the model itself: where determinism fails or where the complexity of the world cannot be learned by the model. Nevertheless, a developed world model represents the most rational and capable analytical instrument available, and can be used at any time for an adaptive, multi-level, and simultaneous policy process.

It is no cure-all. Dilemmas such as the national competition dynamic (“race to the bottom”) on tax, wage, and environmental standards will not disappear through intelligent and rational analysis. Within the operational framework of nation-state action, however, it remains the best option.

Advantages

• Absolutely rational — acts according to moral principles and weighs costs
• Data processing — can integrate the most diverse arguments, scenarios, and information
• Transparency — can disclose costs and trade-offs at any time
• Real-time — can re-evaluate at any moment and identify the best response to a new challenge (war, crisis, etc.) within seconds to minutes
• Inclusive — can incorporate the interests of animals and future generations into decision-making (cost function)
• No conflicts of interest — since the world model has no interests of its own and thinks only of the common good, it is free from conflicts of interest and corruption

Challenges

1. Training a comprehensive world model that is highly capable and not itself biased
2. Designing a suitable cost function (possibly a model in its own right) that learns and understands human interests (justice, security, freedom, affordable housing …). The cost function is therefore at the core of good policy. A kind of democratic cost function could be constructed, in which citizens could decide on outcomes. The system would thus benefit from greater transparency. Alternatively — and in view of cognitive biases and exclusivity — concrete interests could be simulated and measured with extreme precision (discomfort at every moment of a given scenario).

The bottleneck lies not in policy quality, but in the acceptance of “correct” policy. Citizens have grown too accustomed to comfortable, popular politics. To secure acceptance nonetheless, the following principles should be implemented:

1. Gradual Integration

World models could initially be integrated into political decision-making processes as open analytical tools, without immediately replacing existing democratic institutions.

2. Radical Transparency

Citizens should be able to simulate political scenarios themselves, enabling them to understand the consequences and trade-offs of political decisions.

3. Participatory Goal Definition

Societal preferences can feed into the design of the cost function, connecting democratic influence with analytical decision-making.

4. Acceptance as a System Condition

Even the rationally best policy fails without societal legitimacy. Societal acceptance — as well as the potential loss thereof — should therefore be incorporated as a cost factor in the optimisation of political decisions.

5. Normative Anchoring

For such a system to be accepted in the long run, it must be embedded in societal narratives, values, and beliefs that recognise rational and long-term governance as a shared ideal.

Many will probably think that this concept of governance is authoritarian and therefore bad. The belief that democracy is the “best solution” — as Winston Churchill once said: “Democracy is the worst form of government, except for all the others that have been tried from time to time” — proves no longer tenable in the face of existential crises. Obvious cognitive biases, structural problems, and the exclusivity of the right to vote have brought us to where we are today. On the edge of the abyss.