Engineered by slow time, living in (too) fast time
Can we adapt at the speed we create? Some reflections after watching the launch of Artemis II
If you find value in #ComplexityThoughts, consider helping it grow by subscribing and sharing it with friends, colleagues or on social media. Your support makes a real difference.
→ Don’t miss the podcast version of this post: click on “Spotify/Apple Podcast” above!
The problem is not simply that technology is changing fast, but that a species shaped by slow biological evolution now lives inside a system whose rate of change is increasingly generated from within.
Sixty-six million years ago, after the asteroid impact, Earth entered a phase of abrupt ecological disruption. Non-avian dinosaurs disappeared and the future belonged, for a time, not to the strongest species but to those able to survive collapse. Early mammals (ie, our ancestors) were not the rulers of that world: they were small, mostly marginal, and ecologically constrained. Their success was not a sign of superiority in any grand sense, it was rather a sign of robustness under perturbation: the capacity to persist when the system around them was violently reorganized.
That point matters because it reframes the story we like to tell about ourselves. The arc from tiny mammalian survivors1 to a spacefaring species is often narrated as progress. But from the perspective of a complexity scientist, the deeper issue is different: not only what emerged from evolution, but how the tempo of change itself has shifted. What took biology millions of years to produce is now generating technological environments that can transform in years, months or even minutes.
This is a story of acceleration, not just a story of progress.
A tale of biological evolution
On the one hand, biological evolution adapts through mutation, selection, and generational turnover; even under strong pressure, that is a slow process over evolutionary time scales.
I think it is worth spending one minute about this result23, since it is not evident about what time scales we are talking about. From the times of Antropoide we needed about 10 million years to evolve our current brain4 (see also this paper for evolutionary–developmental dynamics of hominin brain size).
If we consider the entire evolutionary timeline from vertebrates to Homo sapiens, the emergence of human language can be interpreted as the outcome of a long process marked by critical phase transitions rather than gradual change alone. A recent study argues that brain evolution crossed successive capacity thresholds through two major transitions: the appearance of the primate cortex with expanded and differentiated functional areas, and the later shift from natural selection to triadic niche construction, in which ecological, neural, and cognitive niches co-evolved and accelerated hominid brain expansion. The figure below provides a schematic view of how these transitions prepared the conditions for human language and, by extrapolation, how a possible third transition associated with artificial intelligence may lead to a novel mode of language (check the original paper5):
A tale of technological evolution
On the other hand, technological evolution works differently. It accumulates outside the genome, through language, institutions, media, machines, and now code. Once change becomes informational rather than strictly biological, it can propagate at the speed of communication and scale at the speed allowed by the underlying networks.
The result is a growing mismatch of timescales: the technological layer of human life now changes faster than the cognitive, social and institutional layers that must absorb, regulate and live with its consequences.
The timescale mismatch
Let me be a theoretical physicist for thirty seconds, and let us assume that humans are a multilevel system with mismatched speeds. At one level there is biology: slow, deeply constrained, written into bodies shaped over evolutionary time. At another there is cognition: faster, flexible, capable of learning, but bounded by attention, memory and limited rationality. Above that there is society: institutions, norms and coordination structures that can adapt, but usually more slowly than the technological infrastructures now mediating them. The problem is not merely that these levels are different: it is that they now evolve at incompatible rates.
When the fastest layer in a system changes continuously, while the slower layers update only intermittently, the tension becomes structural. In fact, technology does not wait for institutions to deliberate, for norms to stabilize or for cognition to recalibrate: it propagates first and leaves adaptation to catch up later.
At the individual scale, this appears as cognitive overload. Human attention is limited, memory is selective and decision-making is bounded by time and context. Yet the informational environment in which those capacities operate is increasingly dense, persistent and competitive. However, more signals do not necessarily produce more understanding: beyond a certain point, they generate friction and difficulty in filtering relevance, greater susceptibility to noise, and reduced capacity for reflective judgment.
At the collective scale, the same mismatch appears as coordination lag. Institutions are designed to stabilize expectations, process conflict and translate knowledge into rules, but institutional time is slow by design: deliberation, negotiation and implementation all require time. When technological infrastructures evolve much faster than these processes, governance becomes reactive rather than anticipatory.
The result is a recurring gap between what systems can do and what societies are prepared to manage.
From a complexity perspective, this is a familiar source of fragility. A system remains resilient when its internal mechanisms of adjustment can absorb perturbations without losing coherence. But when the rate of endogenous change exceeds the adaptive capacity of the system’s stabilizing layers, resilience inevitably begins to erode, although what follows is not necessarily collapse. Instead, more often, it is a persistent condition of instability: local failures, recurrent crises, norm erosion and a growing difficulty in maintaining coordination across scales.
But… we have seen this before
The mismatch is not hypothetical. Modern history already contains repeated episodes in which technological and socio-technical capacities expanded faster than the cognitive, institutional and biological layers needed to absorb them safely.
The first case is urban-industrial modernity. Industrialization increased productive capacity and concentrated populations at unprecedented scale, but sanitation, housing reform and public-health governance arrived later. The early outcome was not uniformly improved welfare, but severe health disruption in rapidly growing cities. Improvements in population health did not follow automatically from economic growth, they required a later phase of political and institutional reorganization. It was “power before safety”.
A second case is fossil-fuel industrialization. Dense, cheap energy transformed production, mobility and living standards, but it also created infrastructures whose consequences unfolded over much longer timescales than the institutions meant to govern them. This is the logic of carbon lock-in: once energy systems are built, they commit societies to future emissions even if political preferences change later. Climate change is therefore not only an environmental crisis, it is also a time-lag problem in which the capacity to alter the Earth system scaled faster than the institutional capacity to constrain that alteration.
A third case is the digital information environment. Online platforms allow information to propagate at speeds and scales that exceed individual verification and often outpace institutional response. False news can spread farther, faster, deeper and more broadly than true news, while polarized attention and homophilic network structure amplify cascades and reduce corrective capacity. Beyond excess information, the problem is thus a structural mismatch between propagation dynamics, cognitive filtering and governance capacity.
Seen this way, the present moment is thus not that much unprecedented. What is new is the density of the feedbacks, the speed of recombination and the fact that the system now accelerates increasingly from within. We have faced timescale mismatch before. But we are now facing it in a more connected system, at larger scale and with less time for correction.
From resilience to adaptability
For most of evolutionary history, survival meant resilience: the ability to withstand shocks and continue functioning. That logic made sense in worlds where change was often severe but comparatively infrequent and where the relevant challenge was persistence under disruption.
However, our condition is different, and the challenge now is not only to survive isolated shocks, but to live inside a system that continuously reorganizes itself. Under these conditions, I think that even resilience — the ability to withstand adversity and “bounce back” to a previous state — is no longer enough, and the more relevant property is adaptability: the capacity to update behaviors, revise norms, redesign institutions and reconfigure coordination as conditions change. In a nutshell: the ability to “bounce forward" into new and unfamiliar circumstances.
However, this shift is far from being trivial, since robustness and adaptability are not the same thing, and they do not always point in the same direction. Robust systems resist change, while adaptive systems remain flexible enough to change without disintegrating. The former stabilizes, the latter evolves. In a high-acceleration environment, too much rigidity becomes a liability, but pure flexibility is not a solution either. Systems still need memory, constraints and shared rules, so the real problem is how to preserve coherence while increasing the capacity to adjust.
Biology cannot speed up on demand. Cognition can learn, but only within the limits of attention and overload. Institutions can redesign themselves, but often only after failure has already exposed the mismatch. This means that the decisive variable is increasingly neither raw intelligence nor technological power, but the ability to coordinate adaptation across levels. But how?
If that diagnosis is even approximately correct, and plausibly assuming that technology will keep accelerating, the question is whether our cognitive, social and institutional architectures can develop enough adaptive capacity to remain coherent inside the environments they are now generating.
Evolution took millions of years to produce a species capable of drastically transforming its environment. What happens when that species accelerates the environment faster than it can adapt to it?
The mismatch is not a detail, it is the problem.
→ Please, remind that if you find value in #ComplexityThoughts, you might consider helping it grow by subscribing, or by sharing it with friends, colleagues or on social media. See also this post to learn more about this space.
I don’t want to overwhelm you with infographics, but this one from Scientific American is really better than a thousand words:
Original paper:
An alternative view of panel b can be seen in the plot below, where instead of showing brain mass vs body mass, there is white matter volume vs gray matter volume:
This is a sharp framing, and the timescale mismatch you describe is real. But I'd push gently on one implication that runs through the piece — that the acceleration of information output is itself the core problem. I think the problem is better understood as a filtering problem, and the good news is that we already have a model for the solution. It's built into us.
Our own neurobiology doesn't attempt to process all available information. It would be catastrophic if it did. Perception is fundamentally selective — sensory gating, attentional filtering, predictive coding — these aren't limitations, they're the architecture that makes coherent experience possible in the first place. The brain doesn't solve information overload by slowing the world down. It solves it by building better filters. The question for the information age isn't how to decelerate technological output — that isn't going to happen. The question is how to design filtering architectures that match the complexity of what we're now producing.
And here's the thing: democracy already operates on this principle. We don't ask every citizen to become an expert in trade policy, epidemiology, and constitutional law. We build institutions — representative bodies, expert advisory structures, deliberative processes — that filter complex information into forms that enable collective decision-making. That's what self-governance actually looks like in practice: not the unmediated exposure of every citizen to every signal, but the structured channeling of relevant information to the appropriate forums for deliberation. The problem you identify isn't that we lack the concept. It's that our current filtering institutions were designed for a slower information environment and haven't been redesigned for this one.
This is where I think AI tools for deliberation become genuinely promising — not as replacements for human judgment, but as the next generation of filters. Tools that can help route complex information to the right deliberative channels, surface relevant context, and support the kind of structured collective reasoning that democracy requires but that no individual citizen can perform alone across every domain. The design challenge is real, but it's an engineering problem with democratic precedent, not an unprecedented crisis.
The deeper prerequisite, though, isn't technological. It's civic. Filtering systems only work when the people participating in them share a basic trust that their fellow citizens have a vested interest in the health of the whole. Deliberation assumes good faith. The institutions we build — whether legislative bodies or AI-assisted deliberative platforms — are only as functional as the underlying commitment of the participants to the project of self-governance itself. That trust is the slow variable that no technology can accelerate, and it's the one that matters most.