Loonshots

Safi Bahcall, a physicist turned biotech entrepreneur, argues that the science of phase transitions, the study of how systems suddenly shift between states (as when water freezes into ice), offers a framework for understanding why good teams kill great ideas and how organizations can nurture what he calls "loonshots": neglected, widely dismissed projects whose champions are often written off as unhinged. Drawing on his experience cofounding a biotech company developing cancer drugs and advising the US government on research policy, Bahcall builds his case across three parts.

The book opens with a personal story. In 2003, Bahcall visited a cancer patient named Alex, the first person to receive a drug developed by his company. Two years later, Bahcall's father was diagnosed with leukemia for which no new treatments existed. The contrast drives his central question: Why do large organizations with excellent people and vast resources so often bury promising ideas? He introduces Richard Miller, an oncologist and CEO of the struggling biotech company Pharmacyclics, who championed a drug called ibrutinib that every major pharmaceutical company dismissed as too dangerous. Miller lost a boardroom battle and resigned, but the clinical trial continued, producing a nearly tenfold higher response rate that led to FDA approval and a $21 billion acquisition.

Part One illustrates what Bahcall calls the Bush-Vail rules, organizational principles named for two innovators. The first story centers on Vannevar Bush, dean of engineering at MIT, who recognized that the US military's weak link was not a shortage of new ideas but the failure to transfer them to the field. In 1940, Bush secured President Roosevelt's approval for a new civilian organization, the Office of Scientific Research and Development (OSRD). Bush separated the loonshot group (scientists exploring radical ideas) from the franchise group (the military fighting the war) while maintaining continuous exchange, a structure Bahcall likens to phase separation at the freezing point of water. The OSRD's most dramatic success was microwave radar: By 1943, B-24 bombers equipped with this technology hunted U-boats across the Atlantic, reducing Allied shipping losses by 95 percent in 90 days. Bush's postwar report, Science: The Endless Frontier, shaped the US national research system. Bahcall traces Bush's approach back to Theodore Vail, who in 1907 created a quarantined research group at AT&T, eventually called Bell Labs, that produced the transistor and eight Nobel Prizes. From these leaders Bahcall derives his first two rules: separate the phases and create dynamic equilibrium between them.

The second story follows Akira Endo, a Japanese microbiologist at the conglomerate Sankyo, who in 1972 discovered mevastatin, the original statin, from a blue-green mold in Kyoto. Endo's drug survived what Bahcall calls the Three Deaths: The medical consensus that cholesterol-lowering was useless crushed interest; the drug failed in rat models (a "False Fail," since rats have mostly HDL cholesterol and very little of the LDL that statins target); and high doses appeared to cause cancer in dogs (another False Fail, later shown to be harmless). After Endo left Sankyo, the company terminated the program and handed the opportunity to Merck. Cumulative statin sales have exceeded $300 billion. From this and the parallel story of surgeon Judah Folkman of Boston Children's Hospital, whose 32-year campaign to block tumor blood supply was validated by the drug Avastin, Bahcall draws practical lessons: Beware the False Fail, appoint project champions to bridge the gap between inventors and operators, and practice LSC, or Listen to the Suck with Curiosity.

The third and fourth stories introduce Bahcall's distinction between P-type loonshots (breakthroughs in product or technology) and S-type loonshots (breakthroughs in strategy or business model). Juan Trippe, founder of Pan Am, was a master P-type innovator who built the world's largest airline through ever-bigger, faster planes culminating in the Boeing 747. Bob Crandall, CEO of American Airlines, was a master S-type innovator whose inventions, including the Sabre computerized reservation system and yield management techniques (adjusting ticket pricing based on booking patterns to maximize revenue per seat), saved American from the devastation of airline deregulation in 1978. Pan Am, blind to S-type loonshots, ceased operations in 1991. Edwin Land of Polaroid followed a similar arc, dominating instant photography for 30 years before his final P-type loonshot, Polavision (instant-print movies), proved a technical marvel with no commercial appeal. Declassified documents reveal that Land understood digital photography better than anyone, having advised President Nixon on digital spy satellites. Yet he dismissed digital for Polaroid because it eliminated film, the company's primary revenue source. Both Trippe and Land fell into what Bahcall calls the Moses Trap: When an all-powerful leader becomes sole judge of new ideas, the organization cannot adapt.

The career of Steve Jobs, Apple's cofounder, illustrates both the trap and the escape. At NeXT (1988–1993), Jobs built a beautiful, expensive machine with no customers. But the animation team at Pixar, which he had acquired and tried to shut down, produced Toy Story (1995), the first fully computer-generated feature film. Ed Catmull, who led Pixar, saw his job as minding the system rather than managing projects, an approach Bahcall connects to his third rule: Spread a system mindset, meaning leaders should analyze the decision-making process behind outcomes rather than just the outcomes themselves. When Jobs returned to Apple in 1997, he embraced S-type loonshots like the iTunes store, separated his loonshot group, and learned to balance it against the franchise.

Part Two explains the underlying science. Bahcall describes phase transitions in traffic flow, forest fires, and terror networks, showing that a tug-of-war between two competing forces determines when each system snaps. He then applies this to organizations. As a group grows, every employee faces a choice between investing in project work and investing in politics. At small companies, individual stakes make project work the rational choice. At large companies, the calculus reverses because any one project barely moves the needle while promotion yields significant rewards. This sudden shift is a phase transition. Bahcall derives an equation for the critical group size, which he calls the magic number: M ≈ (E × S² × F) / G, where E is equity fraction, S is management span, F is organizational fitness, and G is the salary growth rate from promotions. Typical values yield roughly 150, consistent with real-world observations but derived from incentive structure rather than brain volume.

The fourth rule, "raise the magic number," follows from the equation. The Defense Advanced Research Projects Agency (DARPA), created after Sputnik in 1957, eliminates career ladders entirely: Program managers serve fixed terms, reducing the return on politics. Other levers include increasing project-skill fit, shifting rewards from rank to results, and widening management spans. Bahcall proposes that companies appoint a "chief incentives officer" to fine-tune these systems.

Part Three extends the framework to nations. Bahcall addresses the Needham Question: Why did the Scientific Revolution occur in Western Europe rather than in the far larger empires of China, India, or the Islamic world? He compares Tycho Brahe, the Danish astronomer who lost his patron but could move among Europe's many independent rulers until landing in Prague (where he hired Johannes Kepler), with Shen Kuo, an 11th-century Chinese astronomer who had nowhere to go after losing imperial favor because a single emperor monopolized astronomy. Western Europe's fragmented landscape functioned as a loonshot nursery for nations. England specifically succeeded because the Royal Society of London (founded in 1660) established the earliest successful loonshot nursery within a single country, producing discoveries that led to Thomas Newcomen's steam engine in 1712, the technology that powered Britain's global rise.

The book concludes by distinguishing loonshots from Clayton Christensen's concept of "disruptive innovation." Bahcall argues that many transformative ideas, including the transistor, Google's PageRank algorithm, Walmart's rural locations, and IKEA's self-assembly model, began as what Christensen would classify as sustaining innovations (improvements that extend existing products or markets) or as accidents rather than deliberate disruptions. The relevant distinction, Bahcall maintains, is between loonshots and franchises: Organizations need systems to nurture ideas that challenge deeply held beliefs without sacrificing the projects that keep them alive.