Loonshots

by Safi Bahcall

Phase Transitions

Companies behave like materials undergoing phase transitions. The central tension comes down to stake versus rank. Water molecules shift between ice and liquid, same elements but totally different behavior. This is what kills great companies at scale. They freeze.

Nokia sold rubber boots and toilet paper before becoming Europe’s telecom giant. No hierarchy, open mindset. Then they shot down touchscreens and app stores, and Apple walked through that door. Merck discovered Mectozane for river blindness and gave it away free to African nations. They even treated the rivers, killing the parasites without harming the crabs. America’s most admired company before Apple. Then they missed genetic engineering entirely while everyone else moved forward.

How Bush Won the War

Germany was ahead by 1940. Submarines, nuclear fission, Enigma. We were behind and running out of time.

Vannevar Bush left MIT to run the Carnegie Foundation in DC. He got ten minutes with FDR and pitched his plan to close the tech gap. FDR approved on the spot. The Office of Scientific Research and Development was born, the original loonshot incubator. The concept of dynamic equilibrium between water molecules in ice and liquid wasn’t just metaphor. It was the actual operating principle.

Young and Taylor had already shown you could detect ships with sound waves across the Potomac. Volume doubled, went silent, doubled again as waves interfered. Same principle worked for aircraft. The Navy rejected it. Meanwhile in England, Churchill was pushing hard for radar. When the Luftwaffe made their first major attack, radar detected all 144 planes. They shot them down. Germany’s first real loss. Churchill immediately sent a mission to share the technology with us.

Alfred Loomis, a Wall Street guy turned inventor, assembled a team at MIT for microwave signaling. Side note: active versus passive listening in sonar matters. Once radar hit Allied aircraft, U-boats lost their stealth advantage in three months. Allied ship damage dropped to nothing. Admiral Doenitz called it and lost the Battle of the Atlantic. The buzz bombs, those expensive German weapons, met the same fate. Radar spotted them, shot them down.

OSRD did more than radar though. Disease control, plasma transfusions, saving thousands on the frontlines. And then the Manhattan Project, the biggest loonshot of them all.

Bush was scared science would disappear after the war ended. Working with FDR and later Truman, he built America’s research system. It stuck. On the corporate side, Bell Labs became AT&T and built a research division for loonshots. They invented the vacuum tube (think of controlling falling rocks with a wedge to make Morse code) which connected the country through quality landlines. Then transistors, lasers, UNIX, C language. Bush changed government research. Theodore Vail and Frank Jewett at MIT changed corporate research.

Bush-Vail Balance

The key is maintaining dynamic equilibrium between soldiers and creatives. It requires mutual respect from both sides. Bush got things done because he respected the soldiers even while pushing for radical innovation. Jobs didn’t understand this at first, which is why he got exiled. When he came back, he’d learned to respect both sides. Cook is a soldier. Wozniak was pure creative. Jobs’s second act worked because he honored both.

Strict hierarchy kills research teams. Look at 3M’s decline. But armies and assembly lines need that same discipline for maximum efficiency. It’s the difference between creations and franchise operations. The weakest link in most organizations is the transfer between these two domains. Ed Catmull at Pixar, coming from Utah, understood this. He held onto the graphics technology and used Toy Story as a lifeline, avoiding what Bahcall calls the Moses trap: seeing the promised land but never reaching it.

P-Type and S-Type

There are two kinds of loonshots, and most people only pay attention to one. P-type loonshots are product innovations. They make headlines with big technological leaps. S-type loonshots are strategy innovations. They get revealed quietly, often years after they’ve already won.

Pan Am vs American Airlines

Juan Trippe built the Pan Am empire through P-type engineering. He pioneered jet aviation and everyone noticed. Robert Crandall built American Airlines differently. He restructured wages, invented frequent flyer miles, built Sabre ticketing with smart placement, and mastered pricing algorithms for revenue optimization. Nobody paid attention until it was too late.

When government deregulation hit the airline industry, it opened the door for S-type entrepreneurs. Pan Am collapsed. American thrived. Edwin Land at Polaroid tells a similar story. LCD physics, instant photography, 3D imaging, experimental movies. All P-type innovations that dazzled everyone. But he completely missed digital photography. Polaroid went bankrupt.

Pharma Stories

Statins

FDR’s death from heart disease opened the funding floodgates. The Framingham Heart Study launched, the largest population study ever conducted. Researchers raced to solve cholesterol.

Akira Endo, a fungi expert, came to the US with an evolutionary insight. Bacteria and fungi compete constantly. That’s why penicillin (a fungus) breaks down bacterial cell walls. Maybe fungi evolved something to mess with cholesterol too. He found Penicillium citrinum on a rice grain in a small store in Japan.

The problem was other drugs had tried this approach and failed badly. His idea seemed played out. When he tested the compound on rats, nothing happened. The project should have died there. Someone tried chickens by accident and it worked. Turns out rats produce mostly good cholesterol (HDL) while chickens, like humans, make both good and bad (HDL and LDL). Think about how close that discovery came to never happening.

The drug worked on patient S.S. in Japan and looked promising. Then a study claimed it caused cancer. The rumor spread and killed the product overnight. Two American scientists spent months proving the cancer link was a false positive. They fought to bring it back to market as Merck’s Mevacor. There was some prior connection between Endo and Merck, but that doesn’t really matter. A life-saving drug had been rejected for years before finally breaking through.

Avastin

Judah Folkman had a simple idea: starve the tumor by blocking communication between the tumor cell and its environment. The concept was rejected repeatedly for 32 years. When another researcher finally recognized its merit, the company’s valuation jumped $38 million almost overnight.

Others

Ibrutinib behaves like a piranha, latching onto hyperactive cancer proteins and restoring order. It seemed crazy until the FDA approved it. Amgen, working with Professor Eugene Goldwasser, created EPO and then sued every other company trying to produce it.

False Fails

When Friendster collapsed, everyone concluded that social networks simply didn’t work. Peter Thiel looked closer. He investigated the failure and realized the real problem wasn’t the concept. It was infrastructure. Friendster couldn’t handle scale, so the site became painfully slow and users left.

Thiel invested $500k in Facebook. He eventually sold his stake for over a billion dollars. The lesson is straightforward: learn to classify failures correctly. The distinction between a false fail and a true fail is worth everything.

System Mindset

Kasparov’s framework makes an important distinction. Outcome mindset asks why a specific result occurred. System mindset asks what created the conditions for that outcome in the first place.

Traffic jams illustrate this well. They’re emergent systems in the Robert Boyle sense, not governed by simple Newtonian laws. Multiple agents acting and reacting create two distinct states: smooth flow and jammed. Once you hit a critical density threshold, the jam state becomes inevitable. There’s a phase transition between average car distance and average speed. This explains why lowering speed limits in high-traffic areas actually reduces jams. “Ramp metering” and “truck overtaking bans” help maintain the smooth state by controlling density. Economic systems work the same way. So do companies.

Incentive Structures

Consider the cognitive bias in medicine. Surgeons choose C-sections over natural delivery partly because they can bill more. The incentive structure shapes behavior in ways that aren’t always obvious.

This is why companies need someone thinking strategically about incentive design. Moonshot teams require large teams with small salary differences and lots of credit exposure. You want to minimize rank-based competition when people are working on radical innovation. Franchise teams are the opposite. They need small teams competing for speed, where excellence comes from incentivizing by rank. Get the incentive structure wrong and you’ll kill your best ideas before they have a chance to prove themselves.

Needham Question

The Needham Question asks why the scientific revolution didn’t happen in China when virtually all the key inventions existed there first.

The answer comes down to phase separation, or the lack of it. Scholars in India and China were ahead of Europe technically, but emperors and rulers made all the decisions. There was no separation between creatives and soldiers, no protected space for loonshots to develop independently. Shen in China was developing promising insights but got ousted from the empire because he threatened the power structure.

England created the Royal Society of London and gave intellectuals a platform to work independently from political interference. They achieved phase separation and built what became the essential loonshot nursery. The model resembles how Hollywood works today, with production separated from distribution. Or how pharma operates, with research labs independent from sales organizations.

Robert Boyle’s assistant Denis Papin added a piston to an air pump and noted the innovation on the back of a cooking manual. Thomas Newcomen happened to see it and used the principle to build the first steam engine. These kinds of accidental discoveries and cross-pollinations only happen when you create the structural conditions for them. That’s how empires get built.