AHuman engineers often borrow from nature, but the most useful lesson is not always the visible shape of an organism. Animal structures are built under constraints: limited energy, locally available materials and no central architect. Nests, burrows and hives therefore interest designers who want buildings that regulate temperature, move air or use material efficiently.

BTermite mounds are a famous example, although popular accounts sometimes oversimplify them. In several species, the mound is not a giant air-conditioning machine holding one perfectly constant temperature. Instead, a network of channels interacts with daily changes in wind and heat. Air moves as temperature differences develop between outer walls and internal passages. Engineers studying these systems focus less on copying the mound’s appearance than on understanding how geometry and environmental variation create airflow.

CPrairie-dog burrows illustrate another principle. Different tunnel openings can experience different air pressures when wind passes over them. The resulting pressure difference encourages air to move through the underground system. Researchers have tested whether buildings can use similarly positioned openings to support passive ventilation. The challenge is scale: a mechanism effective in a small burrow may behave differently in a multi-storey structure exposed to unpredictable urban winds.

DWeaver birds demonstrate efficient construction with flexible materials. A bird cannot manufacture glue or cut branches to standard lengths; it selects fibres, bends them and locks them through repeated loops. The strength of the nest emerges from many small connections rather than one rigid frame. This has inspired experiments with robotic construction in which machines weave fibres into lightweight shells that use material only where forces require it.

EHoneybees are associated with hexagonal comb, often praised as the mathematically ideal way to divide a surface. The story is subtler. Bees initially build with warm wax that can deform, and the final regular geometry reflects both behaviour and the physical properties of the material. For designers, the lesson is not simply “use hexagons”. It is that fabrication processes and material behaviour can help generate efficient forms rather than forcing every dimension in advance.

FSome animals solve climate problems by choosing when rather than how to build. Desert ants, for example, adjust activity to extreme temperature windows, while certain insects construct nests in orientations that change how much solar heat they receive. Architects interested in adaptive buildings have drawn from this broader idea: a structure can respond to time-dependent conditions instead of being designed for one permanent state.

GBiomimicry also has limits. Natural selection does not produce universal engineering solutions; it produces traits that are good enough for particular organisms in particular environments. A termite mound evolved for termites, not office workers. Human buildings face fire codes, maintenance schedules, costs and expectations of comfort that animals do not. Copying a biological feature without understanding the original context can therefore create an elegant story and a poor building.

HThe strongest projects use biology as a source of questions rather than a catalogue of ready-made answers. How can air move without fans? How can a structure become strong through many weak components? Can a material shape itself during fabrication? These questions encourage designers to test underlying mechanisms. Successful biomimicry is less about making architecture look natural than about learning how living systems work under constraint. The process also changes how engineers define success. A biological system is rarely optimised for a single variable; it balances competing demands such as strength, repair, reproduction and energy use. Human design teams may initially search for the one most efficient natural solution, then discover that resilience comes from compromise. Redundancy, for example, can appear wasteful in a narrow calculation but valuable when part of a structure is damaged. Studying organisms can therefore challenge the assumption that minimum material or maximum efficiency should always be the only objective. Another lesson concerns maintenance. Animal structures are continuously repaired, extended or abandoned as conditions change. Conventional buildings, by contrast, are often designed as finished objects and then maintained reactively. Researchers in adaptive construction are exploring components that can be replaced locally, allowing a structure to change gradually rather than requiring complete reconstruction after damage or a change of use.

APublic debate often treats expert disagreement as evidence that expertise has failed. If scientists, economists or engineers reach different conclusions, observers may assume that the field lacks reliable knowledge. But disagreement can arise for several reasons, and understanding those reasons is essential before deciding whether conflict should reduce confidence.

BOne source is different evidence. Two research groups may be working with datasets collected in different places or time periods. Their methods can both be competent while producing results that are not identical. As more evidence accumulates, the apparent contradiction may disappear because researchers learn the conditions under which each result holds.

CA second source is different models. Experts frequently agree on the observations but disagree about which mechanisms matter most. Climate economists, for example, can use the same projected temperature change yet estimate different financial effects because they make different assumptions about adaptation, technology and the value assigned to future costs. The disagreement is not over the thermometer; it lies in the framework connecting temperature to consequences.

DValues can enter more directly when evidence must be converted into a decision. Suppose two medical specialists agree that a treatment has a small chance of a serious side effect and a larger chance of helping. One may recommend treatment while another advises waiting because they weigh the same risks differently. Their scientific beliefs can be similar even though their recommendations diverge.

EDr Sofia Renn studies what happens when uncertainty is communicated to non-specialists. She finds that audiences often prefer one confident spokesperson to a panel that openly discusses limits. Yet visible disagreement can be informative. If experts identify exactly what they dispute—data quality, model choice or acceptable risk—people can see which parts of the problem are well established and which remain uncertain.

FProfessor Daniel Cho examines consensus. He warns that counting expert opinions is useful only if the experts are sufficiently independent. Ten analysts using the same dataset, software and assumptions do not provide ten independent pieces of evidence. A large majority can therefore overstate confidence when the underlying reasoning is highly correlated. Cho argues that diversity of methods matters alongside the number of people who agree.

GDr Amina Farouk focuses on incentives within institutions. Researchers may be rewarded for novel findings, consultants for decisive recommendations and officials for avoiding visible failure. These pressures can shape which uncertainties are emphasised. Farouk does not claim that experts deliberately distort evidence in every case; rather, she argues that institutional incentives help explain why certain interpretations receive more attention than others.

HDisagreement also changes over time. Early in a research area, multiple explanations may compete because evidence is sparse. Later, some questions become settled while new disputes move to the boundaries of knowledge. A healthy field can therefore contain both strong consensus and active controversy. Looking only at the controversial edge creates the false impression that nothing is known.

IFor decision-makers, the practical response is not to demand the elimination of disagreement. It is to separate questions. What facts do most experts accept? Which conclusions depend on particular assumptions? Where do values enter? What evidence would change the recommendation? A transparent map of disagreement is often more useful than a forced single answer.

JExpertise is valuable precisely because difficult problems contain uncertainty, not because experts can remove it. The quality of advice should be judged by the clarity of reasoning, the evidence used, the handling of alternatives and the willingness to update. Agreement can increase confidence, but disagreement can also improve decisions when it exposes assumptions that would otherwise remain hidden. Media coverage can intensify the problem by selecting disagreement rather than representing its distribution. A programme may invite one supporter and one critic because the format feels balanced, even when one position is held by a small minority of specialists. The resulting symmetry can make a settled question appear evenly divided. Conversely, genuine minority views can be ignored when journalists rely too heavily on institutional consensus. Communicating expertise therefore requires more than counting voices: audiences need some indication of the quality, independence and relevance of the evidence behind them. The same principle applies when experts revise their positions. Outsiders sometimes interpret a change of mind as proof that the original advice was worthless. In science, however, revision after new evidence is often a sign that the process is working. A system in which no expert ever updates would be more suspicious than one in which conclusions become narrower, stronger or occasionally reversed as evidence accumulates.