ANoise is often treated as an unavoidable feature of urban life. Traffic, construction, aircraft, entertainment venues and mechanical equipment all contribute to the acoustic character of cities. Yet environmental psychologists and public-health researchers increasingly argue that sound should be planned with the same seriousness as light, air quality or green space. Their concern is not simply that cities are loud, but that certain patterns of unwanted sound can affect sleep, concentration, stress and people's willingness to use public spaces.

BThe traditional response to urban noise has been to reduce sound levels. Engineers erect barriers beside roads, improve building insulation and impose limits on industrial machinery. These measures can be effective, particularly when a dominant source is easy to identify. But decibel reduction alone does not fully explain how people experience an environment. Two locations with similar measured sound levels may be judged very differently if one contains intermittent engines and alarms while the other is dominated by water, conversation and birdsong.

CThis observation contributed to the development of the idea of the "soundscape". Rather than treating every sound as a pollutant, soundscape research examines the total acoustic environment and the way listeners interpret it. Context matters. Music that is welcome in a festival may be intolerable outside a bedroom at midnight. The same fountain that masks road noise in a square might be irritating in a library. Researchers therefore study not only intensity but also timing, meaning, predictability and the listener's sense of control.

DVegetation is frequently proposed as a solution, although its acoustic effects are sometimes exaggerated. A narrow line of trees is usually insufficient to block substantial traffic noise physically; sound waves pass through gaps and around vegetation. Nevertheless, green areas can still improve perceived acoustic quality. Leaves produce their own sounds, birds may become more noticeable, and the visual presence of greenery can change expectations. In other words, trees may alter the experience of sound even when a sound meter records only a small reduction.

EWater features operate in a related way. A fountain does not normally eliminate traffic noise. Instead, its relatively continuous sound can partially mask irregular noises that attract attention. Designers have found that the character of the water sound matters: a gentle flow may be too weak beside a busy road, while a powerful jet may itself become unpleasant. Successful masking therefore requires a balance between the background noise, the location of listeners and the acoustic properties of the feature.

FStreet geometry also shapes the urban soundscape. Tall, parallel building facades can reflect sound repeatedly, creating environments in which traffic noise persists. Courtyards and spaces shielded from major roads can provide quiet sides to buildings even in dense neighbourhoods. Some planners now encourage housing designs in which bedrooms face these protected areas while kitchens or corridors face noisier streets. This does not make the neighbourhood silent, but it can reduce exposure during periods when quiet is particularly important.

GA further complication is fairness. Urban noise is not distributed evenly. Lower-income residents may be more likely to live near major roads, rail lines or industrial sites, while also having fewer resources to install high-quality windows or move elsewhere. Noise policy that focuses only on city-wide averages can therefore conceal substantial differences between neighbourhoods. Mapping exposure at a fine scale has become an important tool for identifying communities that experience a disproportionate burden.

HTechnology is changing how such maps are produced. Traditional noise maps rely heavily on traffic counts, models and occasional professional measurements. Networks of inexpensive sensors can now collect data continuously across many locations. Smartphones may also contribute, although their microphones vary and measurements made by untrained users can be inconsistent. Researchers are experimenting with ways to combine these imperfect but abundant data with more accurate reference instruments.

IThe aim of contemporary acoustic planning is not to create silent cities. Silence itself may feel uncomfortable or unrealistic in active public spaces. Instead, the objective is to reduce harmful exposure while preserving or introducing sounds that people find appropriate to the setting. This requires collaboration among acousticians, architects, transport planners, landscape designers and residents. A successful city, from this perspective, is not one that has no sound, but one whose sound is considered as deliberately as its physical form.

AForecasting technological change has a poor reputation, and not without reason. Histories of innovation are filled with confident predictions that proved spectacularly wrong: machines expected to transform daily life that disappeared within a decade, and inventions dismissed as curiosities that later became essential. Yet governments, companies and investors cannot simply refuse to think about the future. Decisions about infrastructure, research funding and education often have consequences lasting for decades. The difficulty is therefore not whether to forecast innovation, but how to do so without mistaking uncertainty for ignorance.

BOne common error is to imagine technological progress as a smooth continuation of recent trends. If computing power has increased rapidly, forecasters may extend the same rate into the future. Such extrapolation is attractive because it produces clear numerical projections. However, technologies rarely develop in isolation. Progress may slow because of physical limits, shortages of specialised materials, regulation, energy costs or lack of consumer demand. Conversely, an advance in one field can unexpectedly accelerate another. A prediction based on a single curve can therefore be precise while still being fundamentally misleading.

CEconomist Nathan Rosenberg emphasised another difficulty: the uses of a technology are often discovered after the technology itself appears. Early computers were built for specialised calculation, yet their later applications transformed communication, entertainment and commerce. The economic importance of an invention may consequently depend less on its original purpose than on complementary innovations and new practices that emerge around it. This makes forecasting particularly difficult at the moment when a technology is still novel and its range of uses is narrow.

DThe historian of technology David Edgerton has criticised what he calls an excessive focus on invention. Societies, he argues, are shaped not only by new devices but by the continued use, repair and adaptation of older ones. A dramatic prototype attracts attention, while an existing technology quietly used by millions may receive little notice. Forecasts that concentrate only on what has just been invented can therefore exaggerate the speed with which older systems disappear. Horses, bicycles, paper records and mechanical tools all remained important long after supposedly superior alternatives became available.

EAdoption depends on infrastructure as well as technical performance. An electric vehicle is more useful when charging is convenient; a communication platform becomes more valuable when other people use it; a medical device may require trained staff and reimbursement systems. These surrounding conditions create path dependence: choices made earlier influence which later options are practical. A technically excellent product can fail if introduced into an ecosystem that is not prepared to support it.

FSociologist Everett Rogers approached the problem from the behaviour of adopters. His diffusion model divided users into categories, from innovators and early adopters to the early majority, late majority and laggards. The model is sometimes misused as a timetable, as though every successful technology must move predictably through these groups. Rogers' more important point was that adoption is a social process. People observe one another, assess compatibility with existing habits and respond to perceived advantages rather than engineering specifications alone.

GThis helps explain why expert forecasts can fail even when the experts understand the technology well. Engineers may accurately describe what a system can do but misjudge whether users will accept changes in cost, convenience, privacy or behaviour. Consumers, meanwhile, cannot express preferences for possibilities they have never experienced. Market surveys conducted before an unfamiliar product exists may therefore offer false confidence. Both expert judgment and consumer research contain information, but neither provides a direct view of the future.

HSome organisations respond by using scenarios rather than single predictions. Scenario planning does not ask which precise future will occur. Instead, teams construct several internally consistent futures based on different assumptions—for example, cheap energy versus expensive energy, strict regulation versus permissive regulation, or rapid adoption versus public resistance. The value lies in testing whether a strategy remains sensible across more than one plausible environment. Critics note that scenarios can become elaborate stories with little analytical discipline, but supporters argue that this is a failure of execution rather than of the method itself.

IPrediction markets offer another approach. Participants buy and sell contracts whose value depends on whether an event occurs, and the resulting price is interpreted as a collective estimate of probability. Because participants can profit from being correct, markets may aggregate dispersed information more effectively than meetings dominated by senior voices. Researcher Philip Tetlock has also shown that groups of carefully selected "superforecasters" can outperform ordinary prediction methods on many geopolitical questions. Their success is associated less with secret information than with frequent updating, probabilistic thinking and willingness to change their minds.

JWhether these lessons transfer fully to technological innovation is uncertain. Geopolitical events often have clearly defined outcomes within a limited period; the "success" of a technology can be harder to specify. Is a technology successful when it becomes technically feasible, commercially profitable, widely adopted or socially transformative? Different definitions produce different forecasts. Moreover, technologies interact over long periods, making feedback effects difficult to capture in short-term prediction exercises.

KA further problem is that forecasts can influence the future they claim merely to describe. A widely publicised prediction of rapid growth may attract investment, talent and political support, helping the technology advance. A gloomy forecast may discourage those same resources. In financial markets this type of feedback is familiar, but technological expectations can operate similarly. Forecasting is therefore sometimes performative: belief in an outcome can alter the probability of that outcome.

LThe appropriate response is not to abandon forecasting but to change what is expected from it. A useful forecast need not identify the winning technology twenty years in advance. It can reveal assumptions, expose dependencies and identify decisions that remain robust under uncertainty. The best forecasters are often distinguished not by confidence but by calibration—the ability to match the strength of their claims to the quality of the available evidence. In innovation, humility is not the opposite of expertise; it may be one of its most valuable forms.