10th of May, 2025
The most important problem of the hour, safe artificial intelligence, shares some properties with the development of artificial fertilizers. Here we read from the history of the latter, to inform writing the history of the former.
• • •
Research has two components. We are, on the one hand, trying to find the answer to problems we care about for their own sake. The contemporary archetypal figure for this might be Grigori Perelman, who, despite proving the notorious and lovely Poincaré conjecture, has successfully defended himself against any suspicion of utilitarian ambitions by declining large sums of money, the most prestigious prizes on earth, et cetera. Many similar cases of great minds come to mind.
The second component of research involves solving important problems to directly improve wellbeing in the real world — cue yawning and booing from researchers of all fields. Coming from maths myself, I am inclined to join in with the crowd, but let us resist that inclination together to also name a champion for the practical side of things.
Why don't you start, dear reader, and name some famous researchers who solved a big problem because it seemed beneficial to general welfare.
There are not many, but we can find some B-listers who do fit the bill. Alexander Fleming, perhaps, who found antibiotics. That discovery was an accident of course, so perhaps we should discount it a little, but he was trying to work on vaccines, which is at least related. Speaking of — Edward Jenner, the first doctor to systematize vaccines based on folklore describing immunization through cowpox, is maybe the best candidate. Smallpox was one of the most prevalent causes of death in Jenner's time, and he decided not to patent the vaccination to help save lives as quickly as possible. It seems at least likely that his main motivation for starting this line of inquiry was helping people.
But something is not satisfying with Jenner and the smallpox vaccine as our champion for practicality. The discovery appears as too effortless to be relatable to a modern scientist: Put cowpox on patient, wait, put smallpox fluid in patient, no infection, done. Where is the dramatic arc? People hadn't quite warmed up to the concept of solving problems, so the scientific community clamoring for a solution to the indeed very severe problem seems small in retrospect. Searching the archives of the Royal Society for document titles including "pox" in the 18th century yields 25 results, while "fossil" yields 26.
Ideally, we would want to examine the history of a pressing problem that needed significant effort before it was cracked. With lots of money being thrown at it.
• • •
In The Wheat Problem, published 1900, physicist Sir William Crookes analyzes the world's agricultural economy and comes to a bleak conclusion. The limited area available for cultivation of wheat, he forecasts, will only sustain Europe's rapidly increasing population for 20 more years. Within this short time span, catastrophic famine would become inevitable.
Crookes writes, in a tone eerily familiar to those grappling with today's problems:
Many of my statements you may think are of the alarmist order; certainly they are depressing, but they are founded on stubborn facts.
As with most diagnoses of impending doom, it seemed like science might have an answer to the wheat problem:
I hope to point a way out of the colossal dilemma. It is the chemist who must come to the rescue of the threatened communities. It is through the laboratory that starvation may ultimately be turned into plenty.
If there is not enough fertile land to sustain the population at current crop yield, the only way out is to increase the yield, which in the short term means applying more nitrogenous fertilizer. At the time, the most important source of it was bird excrement, called guano. It was mined in Chile, Peru, Namibia, and various islands in the Pacific, and considered so important that the United States passed an act that renders any uninhabited guano-rich island occupied by a citizen an annexed U.S. territory. As one can imagine, shipping off millions of metric tons of bird leftovers quickly depleted the deposits, and instead transporting nitrate-rich Chilean soil was considered the next substitute.
Such a costly (and as Crookes pointed out, ultimately futile) logistical effort meant agony to chemists. The very air they were breathing was three quarters nitrogen, known to be the main ingredient for the formidable fertilizer component ammonia. Even the water they were drinking taunted them with its two thirds hydrogen, the only other ingredient. Existing methods for synthesis like those based on the Frank-Caro process were much too energy-intensive. The game was on: Who could turn air into bread?
Shortly after Crookes raised the problem, chemists got to work. Wilhelm Ostwald, distinguished professor in Riga and Leipzig, studied catalyzers for the ammonia reaction. In his 1903 retrospective Nitrogen — The Mill of Life, he writes about his research:
Great scientific inventions, formerly considered a spontaneous gift of the heavens, can now be achieved via sheer force of systematically organized science. We do not need to wait for our century's genius any longer.
Despite his best efforts, while considering ammonia a "life-or-death-question", he did not find a feasible method of synthesis. First filing a patent about technical iron as a promising catalyzer, he did not realize the presence of nitrides in the iron as a necessary factor, failed to replicate his experiments, and retracted the patent.
The whole affair was complicated by the circumstance that the needed answers in both chemical science and industrial engineering were being developed in parallel. It appeared that simultaneous high-pressure, controlled-temperature synthesis from H₂ and N₂ was the answer, requiring technical stunts to even obtain accurate measurements, not to speak of global-scale production. An upcoming chemist in Karlsruhe, Fritz Haber, was able to demonstrate synthesis nonetheless.
Prospects for feeding the world were crushed once again when established professor and lamp inventor-millionaire Walther Nernst attempted to replicate the less experienced Haber's experiments. From a verbal exchange with Haber at the Bunsen Society meeting, published in the Zeitschrift für Elektrochemie, in translation from Benjamin Johnson's excellent Making Ammonia:
It is regrettable that the equilibrium of ammonia formation is shifted further toward the low end of the yield than one would have supposed from Haber's highly inaccurate numbers. We could have then considered synthetically producing ammonia from hydrogen and nitrogen. But the situation is actually much less advantageous, the yields are roughly three times smaller than expected.
Haber was not deterred. On the contrary, perhaps inspired by his critic's success in commercializing the Nernst lamp, he turned to the German chemistry corporation BASF to further develop his high-pressure hydrogenation project. There he met Carl Bosch. The encounter seems to have been a purely coincidental decree from BASF management, with both having an entirely unimpressive profile up to this point. And yet, in retrospect, Haber had just been paired up with one of the greatest industrialists of his generation. Because Bosch and the team were good at optimizing for any given narrow objective, ammonia synthesis started to look like a well-conditioned agenda; the higher the pressure, the higher the yield.
Their work resulted in the Haber-Bosch process, which started displacing natural fertilizers in 1913, just in time to prevent Crookes' forecast from becoming a reality. Hundreds of millions of tons of ammonia are still made this way each year.
It might pay off to mention that Fritz Haber is not the hero of this story, despite playing the most instrumental role. In World War I, he was also chiefly responsible for the development and weaponized deployment of chlorine gas and phosgene. This idea was cruel even for the political climate of the time, and while up to the present day one finds a tendency to justify the unjustifiable just like Haber claimed chemical weapons to be humane, no such excuses shall be granted here. Carl Bosch on the other hand leaves a more pleasant impression, leaving a legacy of charity and a paper trail in the 30s that could be understood as a good effort to preserve liberal democracy during the rise of fascism.
What lessons can we draw from ammonia? One salient observation is that this pan-European problem was solved entirely in Germany, specifically Prussia. This could be because it was overextended in agricultural production, heavily depending on imports of natural fertilizers, and thus incentivized through higher costs, whereas other world powers did not pay import premiums because they directly exploited natural resources in their colonies. The main contributions to research came through public universities, even though capital expenses for ammonia production were too large to start commercialization as an academic, in contrast to lucrative research projects like the Nernst lamp. Haber simply became a paid employee of BASF. Nernst contracted with the Griesheim-Elektron company in hopes of selling a patent. Reading Ostwald's 1903 publication for instance, similarly to other writing by Nernst and Haber, one gets the impression that they were mainly interested in solving the chemical puzzle, with money serving as a secondary and societal impacts as a tertiary motivation. Making Ammonia devotes a section to Haber's motivation. A highly interesting piece of evidence it cites is a 1903 letter about the Austrian Margulies brothers' plans for industrial ammonia synthesis from Haber to Ostwald:
Esteemed Herr Geheimrat!
An Austrian company has recently written to me repeatedly to hear my opinion on whether it would be worth their while to catalytically produce ammonia from nitrogen and hydrogen on a large scale. I answered by calling their attention to the low costs of producing ammonia as a byproduct of coking plants. This advice did not decrease the company's inclination toward producing ammonia via the above mentioned process. A foreign colleague approached me at a conference in Berlin and said that you, hoch geehrter Herr Geheimrat, have studied this question and have worked out a technically useful method. Therefore, please let me know whether it would suit you if I make the necessary arrangements for the Austrian company to contact you. I am also driven by the thought that you may find it desirable that a financially strong and reputable company take over the technical implementation. [...]
Just as we needed fertilizers to continue increasing the world population without devastating famine, we now need safety research to increase machine intelligence without loss of control. As Crookes, we are looking for prior art, scale, incentives, big customers, and founder-researchers who — for one reason or the other — care to get to the solution.