Prediction isn’t the only — or even the most productive — way to influence what comes next. We cannot predict the far future with meaningful accuracy; the complexity and compounding uncertainty make it practically impossible. We can, however, shape it by defining clear and desirable visions of the future, working backward to understand what conditions must be met, and enabling what moves us in that direction.
Short-term forecasts (economics, technology, etc) fail regularly, while long-term ones collapse under the weight of complexity. The further we look ahead, the more accuracy decays. Philip Tetlock’s research on expert political forecasting found that most long-term predictions perform slightly better than sheer chance once they stretch beyond a few years. This pattern is consistent across domains: weather models diverge, market predictions break down, and even science and technology roadmaps routinely miss where transformative innovation will arise, often predicting steady progress while overlooking disruptive leaps.
Part of this is structural — the future isn’t a fixed path waiting to be discovered but a dynamic system sensitive to all inputs. The other part of it is psychological. Predicting outcomes biases us toward currently legible risks — artificial intelligence, climate change, nuclear war, global pandemics — while neglecting the unknown unknowns that, historically, have shaped the world most profoundly. No one forecast the internet, CRISPR, or the Arab Spring with precision. Our forecasts also skew negative. We are wired to anticipate loss over possibility. This defensive bias may help us survive, but it narrows our imagination of what can be built.
When prediction becomes our dominant mode of engaging with the future, we default to mitigation and risk avoidance — useful in crises, but lacking as a philosophy of progress. Prediction treats the future as something to be avoided, not authored.
People are more inspired by a vision to build toward than by a warning to avoid disaster. Envisioning what a flourishing long-term civilization could look like — stable, abundant, equitable — allows for more broad participation and alignment. It expands imagination beyond immediate risks, opening space for innovation.
This approach begins with backcasting rather than forecasting: define a desirable future state, then work backward to identify the steps, conditions, and enablers that must exist for that vision to be realized. For example, the Apollo Program was not born from a probabilistic model of spaceflight. It began with a clear declaration — to land a human on the moon — and mobilized capability around that goal. Likewise, the United Nations’ Sustainable Development Goals (SDGs) function as vision-based coordination tools, effective at aligning effort and funding around shared ideals rather than predictions of collapse.
Vision-based planning also builds adaptive capacity. When we work from a vision, we can measure progress against whether actions move us closer or farther from it, not whether a specific forecast proved to be correct. This enables continuous course correction without losing sense of purpose. It turns uncertainty from a threat into a feedback mechanism.
To sustain this approach, we must:
This is how vision becomes a method, not just a metaphor: a structured way to act with purpose amid uncertainty.
If prediction tries to make the future legible to the present, then vision makes the present legible to the future. The former seeks control through knowledge; the latter through purpose. Prediction narrows; vision expands. The predictive mindset aims to minimize error, while the visionary mindset seeks to maximize potential. Both have their place — we must still assess risks and plan responsibly, but prediction alone cannot carry the moral and creative weight of shaping the far future.
Longtermism should be guided not by forecasts of doom but by the courage to declare what is worth becoming, and the discipline to work backward until it becomes real.
We cannot know the future, but we can choose to create it.
A very lucid view. The vision is constructed from what we know about our current conditions, so it is more realistic and tailored to our knowledge. Long-term predictions are often erroneous because they ignore logically unforeseeable circumstances (technological and cultural changes).
But we must not fall into the error of constructing a "vision" from entirely contemporary elements. We must know how to extract the essential and promising from the present. Current progressivism, for example, is based on a political model that is probably exhausted. Let us remember that Voltaire and Montesquieu advocated humanist development... but they could not foresee the structural political changes (universal suffrage, political human rights, etc.) that this would entail.