the new enabling tools How long does it take for a new tool to spark a major discovery? The gap in time is getting shorter over time: in the 1800s, it was an average of 30 years; in the first half of the 1900s, it reduced to 21 and then in the second half to 10 years; and it finally fell to 6 years between 2000 and 2022 (as we see in Figure 1.5a). Since 2000, the average time from tool to discovery has narrowed to 5 years in chemistry—and 1.5 years in medicine and biology. Yet this pattern has taken place in a largely improvised, piecemeal way and without a general awareness or theory—without realising that there is a principle of tool innovation behind discovery. As we would expect, the gap in time is shorter for discoverers who create a new tool (seen in the darker bars in Figure 1.5b). This is important: method-inclined researchers see quicker returns and impact. The decreasing delay over time illustrates that science—in part—is becoming more efficient. Advances in new technologies like higher-performance computers, quicker data analytics, high-speed imaging and rapid AI and machine learning methods enable running faster experiments, processing data more quickly and uncovering patterns that would have been impossible a few decades ago. But we are only scratching the surface and there is enormous room to make better tools and discoveries and make them faster and smarter—as we explore in depth in Chapter 6. A key insight emerges here: discoveries generally follow soon after the new key tool is developed. In fact, we find that 10% of discoveries are sparked in the same year— or simultaneously—that the enabling tool is created, 21% emerge within two years, 33% within four years and 53% within ten years. Since 1975, most discoveries— 51%—arise within just four years of the new method innovation. This highlights how innovation in tools is inseparably linked to innovation in science (Appendix Figure 1.9). medicine/biol 05 Avg. number of years Avg. number of years 10 15 17 <1600 20 61 138 233240 26 30 26 21 16 10 7 6 8 <1600–1699 1700–1799 1800–1899 1900–1949 1950–1999 2000–2022 12 14 12 12 10 11 10 4 3 20 0 50 100 150 200 250 All discoveries Discoveries that are a method or tool, or made using one developed by discoverer All discoveries Discoveries that are a method or tool, or made using one developed by discoverer Contemporary science: all nobel-prize discoveries-by field Evolution of science: all major discoveries over time chemistry physics econ/social astronomy Figure 1.5 Gap in years between the developed method or tool and the discovery it enabled—across time and fields The data show science's 761 major discoveries (Figure a)—and all 533 nobel-prize discoveries (Figure b). Here we expand the analysis and combine all nobel-prize and major non-nobel discoveries during the same time period across these five fields (in Figure b); we find similar results: for example, the share of discoveries that either are a method or tool—or are made using one developed by the discoverer—remains consistent at 14%, 12%, 12%, 10% and 3% (left to right). S What patterns emerge across fields? Astronomers are by far the quickest in exploiting new tools: more than half (54%) of nobel-prize discoveries are triggered within two years—and a remarkable 88% within ten years. Take the discovery of a supermassive compact object at the centre of our galaxy, the Milky Way. It was uncovered in 2002 by the German Reinhard Genzel and American Andrea Ghez. The groundbreaking finding was sparked by a new adaptive optics spectrometer constructed the same year, together with the world's largest telescopes. Or consider the discovery of the first exoplanet. It was revealed in 1995 using a new fibre-fed echelle spectrograph built in 1993 at the Haute-Provence Observatory in Southern France. This astonishing breakthrough redefined how we view our own planet—and our very place in the vast universe. Some fields are slower at translating new tools into new breakthroughs. In medicine and biology, 51% of nobel-prize discoveries emerged within 10 years; in economics, it is 54% (Appendix Figure 1.9). But how we can narrow the tool-to-discovery gap across fields is a key question we tackle later. What is one of the most direct signs of the power of our tools in driving science? When the moment a tool is invented, it sparks a breakthrough. Take the spectrograph, created in 1859 at the University of Heidelberg. As soon as it was built, it enabled discovering atomic light signatures: that each chemical element gives off a unique light signature. This groundbreaking instrument allows us to produce properties of light in the electromagnetic spectrum, measure the mass of atoms and molecules and identify what planets and stars are chemically made of. The spectrograph became a cornerstone of modern astronomy, chemistry and physics. Take the microscope designed with a silver staining technique in 1873. It immediately led to discovering the structure of our nervous system—and the first view ever inside our previously invisible nerves and how they could function ( Picture 1.6). Picture 1.6 The first image of a nerve cell, depicted with a new silver staining technique. Reproduced from Wikimedia Commons. The mass spectrograph invented in 1919 immediately sparked the mapping and identification of different isotopes—atoms of the same element that have different 44 T HE ENGINE OF SCIENTIFIC DISCOVERY masses. Some are stable, others radioactive. Isotopes help power medical imaging, cancer treatment, carbon dating and even natural resource exploration. The frequency variation method designed in 1924 immediately made the experimental detection of the ionosphere possible—a layer of our upper atmosphere at the edge of space. It is home to many of our satellites and influences our radio and GPS signals. An atom trap—called the time orbiting potential trap—developed in 1995 along with laser cooling immediately enabled the discovery of an ultra-chilled state of matter: Bose–Einstein condensate. This extreme state of matter allows producing nanotechnology and manipulating matter on an atomic scale. A powerful expression cloning strategy created in 1997 immediately led to discovering our elusive receptors that detect temperature. It revealed how our body translates heat and cold into electrical signals in the brain. This fascinating breakthrough by the American David Julius at the University of California and the Lebanese Ardem Patapoutian at Scripps Research in California uncovered the molecular basis of touch, pain and temperature. These are just a few examples of the fastest route to discovery: creating a tool that instantly reshapes what we can detect, measure or understand. And it is not rare. The immediate method-to-discovery link makes up about one in every eight discoveries since 1975—and about three in every ten discoveries are uncovered within two years. That is not just remarkable, it is key to understanding the nature of discovery(Appendix Figure 1.9). In some fields, it can take a year or longer to run experiments with a new tool, write up the study, go through the peer review process—in one or more journals—and publish the breakthrough findings. Discovery time lags exist across science: defined as lags in making breakthroughs when a new method already exists for years but is not immediately applied to tackle a specific problem. They are important missed opportunities to spark advances using new methods not yet exploited—within and across fields—over longer periods. Take the very innovative technique called optical tweezers—initially created in physics in 1970 and published in the prestigious journal Physical Review Letters. This laserbased tool can grab and hold (like with tweezers) particles using beams of light. But it was not until 1987 that scientists finally applied it in biology and medicine. Suddenly, researchers could manipulate DNA strands, viruses and even living cells with extraordinary precision. Today, optical tweezers are used for everything from studying the structure of proteins to running diagnostic tests for diseases like malaria. Take the discovery of cyclin in 1982—the proteins that control the cell cycle. The discoverer, the British biochemist Tim Hunt, said that 'I did a very simple experiment … It is very surprising that nobody had spotted this before; it could have been done any time after the invention of the slab SDS polyacrylamide gel [electrophoresis] in about 1970' . That breakthrough method for analysing proteins was invented by the Swiss molecular biologist Ulrich Laemmli—published in a landmark article in Nature, Cleavage of structural proteins . It has played an enormous role in helping understand cancer and designing drugs. The tool was ready; but no one had yet used it that way. The same pattern appears again and again across science. In development economics, one of the greatest transformations in understanding the causes of poverty S and low education came from leveraging a unique experimental method: randomised controlled trials (RCTs). Employing the method, Michael Kremer, Abhijit Banerjee and Esther Duflo received the Nobel prize for having 'introduced a new approach to obtaining reliable answers about the best ways to fight global poverty' . Yet the RCT method had long been developed in medicine —with over half a million RCTs already conducted up to 2003, when economists used it for the first time. Economists did not reinvent the RCT method but applied it to economic questions. In fact, the earliest experimental trial is often traced back to 1747, when Scottish physician James Lind proved that oranges and lemons are effective in treating scurvy (a severe lack of vitamin C) among sailors at sea. So why do we have large discovery time lags across science? When there is a delay of more than several decades, the new tools are often developed in other fields and not immediately known in one's own field. Think of optical tweezers and the RCT method already long employed in other fields. Because researchers are rarely trained to look beyond their own field for methods, many important discoveries are unnecessarily delayed. So some discovery delays stem from artificial borders across fields and their methods. Yet methods are still the binding constraint—but disciplinary isolation reinforces that constraint and creates a longer gap. Other discovery delays stem from scientists trained to focus on findings and theories, not always on testing and experimenting with a wider range of tools that enable them. The key is that whether there is a long time lag or not, the main spark of discoveries is generally the new tools never applied before to the specific problem—even if created in one field and adopted later in another. So discovery delays often represent some of the most overlooked opportunities and low-hanging fruit in science. Because new tools can help cure diseases and improve society, every year or decade they lie unused is a year or decade of missed impact. Closing these gaps accelerates breakthroughs—from life-saving treatments to advances in clean energy. Ultimately, none of these transformative discoveries would have been possible without these remarkable tools—hidden in plain sight in other fields (Table 1.1). But with time lags for years—or even decades—for some discoveries, what is then a new tool? A new tool does not just need to be invented, it also has to be used. Innovation comes from application, not just invention. We find that 70% of discoveries in the last half-century are made within ten years of creating the enabling tool—and 90% within twenty years (Appendix Figure 1.9b). Even if a tool was invented decades ago, the first time it is ever applied to a new problem is what commonly unlocks breakthroughs. Yet the longest delays are mostly in the early history of science, when the scientific community was smaller, communication was limited and researchers were scattered across vast distances—significantly slowing progress. Take the telescope. It was invented in 1608 and then improved, yet major breakthroughs like discovering the distance to the sun in 1672, the motion of stars in 1719 and galaxies in 1750 came much later. Even longer time lags occurred before the 1600s. In these exceptional cases, the tool had never been used before to tackle the problem, but with the new lens, the discovery was then possible. Scientific progress relies not just on invention but access: on getting the right tool in our hands.      T able 1.1 Discovery time lags: delays between the new tool and a discovery (that slow real-world impact)—examples across different time periods Making the discovery is within x number of years of developing the method used Method or tool developed → the discovery made using it 1600–1900 1901–1990 1991–2020 pneumatic trough, improved (1774) → oxygen (1774) electron tube, improved (1913) → laws governing impact of electron on an atom (1913) time orbiting potential trap (1995) → Bose–Einstein condensate (1995) Impact of new methods and thus discoveries directly realised0y e a r s (same year) spectrograph, improved (1859) → atomic light signatures (1859) frequency variation method (1924) → ionosphere (1924) spectroscopy, improved adaptive optics (2002) → supermassive compact object at galaxy's centre (2002) microscope, with silver staining technique (1873) → structure of nervous system (1873) electron image processing, improved (1990) → electron cryo-microscopy for bacteriorhodopsin (1990) LIGO detector, improved (2015) → observing gravitational waves (2015) telescope (1608) → Jupiter's moons (1610) starch-column chromatography (1948) → synthesis of polypeptide hormones (1953) 1–5 years microscope, improved (1662) → cells (1665) particle accelerator, Bevatron (1954) → antiproton (1955) neutrino detector, improved fibre-fed echelle spectrograph (1993) → first exoplanet (1995) (1996) → neutrino oscillations (1998) quadrant electrometer, improved (1897) → chemistry of radioactive substances (1899) spark chambers, improved (1959) → symmetry violation in decay of neutral K-mesons (1964) in vivo coimmunoprecipitation assay (2000) → oxygen-sensing mechanism in cells (2001) thermometer, mercury (1714) → Fahrenheit scale (1724) metal-oxide-semiconductor (1960) → CCD sensor (1970) electron density map, improved (1991) → spatial structure of potassium ion channel (1998) Impact of new methods and thus discoveries postponed Time lag 6–10 years leyden jar (1745) → nature of electricity (1752) molecular spectroscopy (1939) → magnetic moment of electron (1947) telescope, Hubble space (1990) → accelerating expansion of universe (1998) galvanometer (1820) → Ohm's Law (1827) ultraviolet spectrophotometer, configurationimproved (1954) → of ribonuclease (1961) microinjection, improved (1991) → RNA interference (1998) battery (1800) → electromagnetism (1820) chromatography (partition) (1941) → carbon dioxide assimilation in plants (1952) laser, titanium sapphire (1982) → stimulated emission depletion microscopy (1994) 11–20 years isomer counting (1875) → linkage of atoms in molecules (1893) transistor (MOSFET) (1960) quantized hall effect (1980) → Northern blotting technique (1977) → toll-gene in immune system (1996) discharge tube (1875) → x-rays (1895) electrophoresis (SDS-PAGE), improved (1970) → cyclin (1982) atomic force microscopy (1986) → production of graphene (2004) eudiometer, improved (1777) → atoms (1803) superconducting quantum interference device (1964) → superconductivity in ceramic materials (1986) … 21–50 +y e a r s galvanometer (1820) → Joule's law (1841) x-ray crystallography (1913) → structure of DNA (1953) … microscope, improved (1826) → antiseptic medicine (1867) ultraviolet spectrophotometer (1941) → crown ethers (1967) … Methods or tools—and the discoveries they enabl ed—are based on science's 761 major discoveries—w ith examples shown by time period and time lag. 48 T HE ENGINE OF SCIENTIFIC DISCOVERY We face time lags in not only generating discoveries but also realising the importance of some. When the Moravian-Austrian monk Gregor Mendel statistically studied peas, his research went largely unnoticed for decades, but eventually contributed to how we understand inheritance and genetics. When the British Charles Babbage and Ada Lovelace designed the first mechanical computer and computer programme, and when the British Ronald Fisher vastly expanded statistical methods, little did we know that most of science—including governments and our everyday lives— would rely so heavily on computers and statistics. We do not always immediately grasp, and cannot yet imagine, the important impact that discoveries eventually trigger. The same goes for the Wright brothers. They tested flying gliders in a small town in North Carolina that eventually paved the way for airplanes and how remarkably connected we are today around the world. Yet such seminal breakthroughs revolutionise our lives and how we understand life itself. Let us look beyond the timing of tools to an important question: have scientists been able to predict some discoveries before they happened? Some discoveries have in fact been anticipated because the right tools and knowledge were already in place. Yet predicted breakthroughs make up only a small fraction of major scientific advances. Take the discovery of DNA's structure. DNA was first isolated in 1869, using chemical extraction methods and new advanced microscopes capable of observing cell nuclei. Once DNA was detected, we knew—and could predict—it must have a structure. And once x-ray crystallography was invented in 1913, we also had the necessary method at our disposal that could reveal molecular shapes. It was only a matter of time. Eventually, researchers used Rosalind Franklin's crucial x-ray images to uncover DNA's double-helix structure—in 1953. A similar story played out with the human genome. After inventing DNA sequencing in 1977, it was no longer a question of if we could decode the entire human genome, but when. A powerful existing method—gel electrophoresis—played a crucial role in separating DNA fragments in the sequencing process. What followed was a competitive race across research teams for finding the best way to sequence the whole genome. The publicly-funded Human Genome Project was launched under geneticists James Watson and Francis Collins. A large privately-funded initiative was led by biotechnologist Craig Venter at his genomics firm. In 2000, President Clinton and Prime Minister Blair intervened, requesting open access to the data to coordinate efforts. An integrated method for genome mapping developed in 2001 sped up the discovery process—and the finish line came into view. It ultimately enabled achieving the incredible feat in 2003: mapping the entire human genome. This milestone transformed modern biomedicine—allowing us to better diagnose and treat diseases—and also led to major advances in forensics and even our understanding of human evolution. A particularly exceptional case is gravitational waves, predicted a year after Einstein published his general theory of relativity in 1915. Gravitational waves were predicted to move through space at the speed of light—and they were, at first, a hypothetical phenomenon, speculative for Einstein himself. Einstein's theory was made possible through advanced mathematical methods and experimental findings, especially by Michelson and Morley on the speed of light in 1887 that applied S the recently invented interferometer. Even those discoveries that at times seem—at first glance—to be largely theoretical breakthroughs rely heavily on recently developed powerful tools (Chapter 6). And later, we also first had to create an instrument sensitive enough to physically detect the waves. A century later, the LIGO detector was constructed—and it was updated in 2015, and within just days, it captured gravitational waves. The detected waves were about 1.3 billion light-years away, meaning the signal had travelled for over a billion years before reaching the earth. Another example is the periodic table of elements—created by the Russian chemist Dmitri Mendeleev in 1869 while working at Saint Petersburg State University. He achieved the breakthrough using an early classification of chemical elements in 1829 and the mathematical law of octaves developed in 1865. He also indirectly relied, beforehand, on several new elements recently identified using the new spectrograph in 1859. And afterwards, the spectrograph was also used to identify some of the predicted elements where empty places remained in the periodic table. It was just a matter of detecting the missing elements we knew existed. The elements, as a whole, make up the world we live in and the foundation of chemistry. So uncovering these exceptional discoveries is often a question of developing or applying the missing method—filling in the missing pieces (as seen in Figure 1.6).