On the campus of the University of Glasgow, passing a 17th-century gate, a dilapidated brick building is the laboratory of the Regis Professor of Chemistry. The situation has not changed much since King George III appointed the first title in 1818. Experiments are still carried out in glass flasks-although they are now carried out by students wearing T-shirts and jeans.
The current Professor Regius wanders the building in a sports tweed jacket and khaki pants, declaring that everything will be different soon. "In any physical or biological laboratory, there is automation," Lee Cronin told me. "In chemistry, all this is still done by hand." Cronin opened the door of an empty room, where the chemical reaction was bubbling under a rudimentary robotic scaffold. He revealed that chemical automation is already in progress and its goals are far. More than industrial efficiency.
Cronin is committed to repositioning chemistry as the science of the 21st century. Since coming to this university as a 29-year-old lecturer in 2002, he has established a 65-member research group, which is one of the largest research groups in the field of chemistry, with an annual budget of nearly US$5 million. About half of these resources are used to develop "chemical computers"-Cronin's peculiar name for computer-driven automated chemical laboratories. In addition to the potential of his chemical computer to customize specialized drugs for personalized medicine, Cronin also hopes to computerize his field of chemistry. He believes that this is the only way to successfully deal with the two prominent challenges in science: discover the origin of life, and advance artificial intelligence by making machines as intelligent as the human brain.
In his view, these issues are related, because life and intelligence are produced from prebiotic chemistry. Finding a chemical transformation from basic substances to Homo sapiens will require more experiments than pouring liquids into flasks with both hands. The scope of his work is so compelling that the Defense Advanced Research Projects Agency (DARPA) supports one of his projects. The Templeton Foundation also provided $2.9 million in grants to Cronin and several colleagues to understand how life started. And he is astute, knowing how to supplement this money by developing practical applications for his chemical computer at the same time.
"He has made great progress in pharmaceuticals and all other fields," said Sara Imari Walker, an astrobiologist at Arizona State University, one of his main collaborators on the origin of life. "What he did well was to strategically use other fields to get the basic science he wanted to accomplish."
If Cronin's ambitions are realized, countless other researchers will strengthen his research by making breakthroughs on their own chemical computers. He is eager to "witness the self version of Moore's Law" through chemistry, which is a phenomenon in which computing power doubles every two years. Although his rhetoric was met with some counterattacks from others in his field who questioned whether automation would bring such a revolution, Cronin was not bothered by these doubts. The computer in the laboratory is not behind us either, ignoring our conversation and wholeheartedly assembling a molecule that few human chemists can manually synthesize.
When Cronin was 8 years old, he rummaged through his parents' house, looking for components to assemble the computer. In order to distract him and save a few surviving electrical appliances, his father bought him a set of chemicals. Cronin immediately set about combining it with the electronic parts he picked up. He doesn't have the idea of a chemical computer in his mind—at least not—but he has already begun the process of combining science and technology, which will determine his life's work as a freelance experimenter-inventor-entrepreneur. "I have always been interested in reality," he said.
At the school in Ipswich, a small town in the east of England, this didn't go well. The education system has a low tolerance for precocious puberty, and his teachers especially dislike students asking questions that the teacher cannot answer. "Everyone said I was too stupid to do what I wanted to do," he recalled. They think that his questioning of their courses is an escape method to avoid the real work of rote memorization. As a result, he became more and more destructive in class, while using his spare time to teach himself relativity mathematics. His grades dropped so low that he was not eligible to take the exams required by the university. His father intervened again. He paid the registration fee for the entrance exam out of his own pocket, and his understanding was that if his son passed, he would be reimbursed. When the results came out, it became clear that Cronin was not actually stupid. He was admitted to York University.
"I spent half of my time dreaming about science and half of my time doing chemistry, which I find boring," Cronin said. To a large extent, his training in chemistry is similar to learning woodworking, mastering a series of chemical reactions that can be used to construct new molecular structures. "I have been thinking,'What is the smallest object that I can assemble into the most complex object and trigger a chain of events?'" he said. In other words, he is not as interested in the details of chemistry as what chemistry can create.
During this period, in the 1990s, Cronin also learned the practical skills needed to support his non-traditional career: obtaining a doctorate. After taking a faculty position at the University of Birmingham, he began to work hard to become the best molecular carpenter. "I didn't give up my philosophy," he said. "I [just] realized that if I didn't become a successful chemist first, I couldn't be a successful scientist." By the time he moved to Glasgow in 2002, he could build almost any molecule on demand.
In order to avoid being constrained by the workbench, he began to try to use a simple robot system to move the liquid. Combining off-the-shelf hardware, basic open source robotics and laboratory equipment, he built machines that can automate experiments. Cronin finally grew up in Glasgow, where he tried to build the machine of his childhood dreams.
Back in the laboratory, I found the doctor. Student Przemyslaw Frei stands in front of the 3D printer, watching its nozzle extrude layers of translucent plastic.
"This is the most complicated reactor I have built," he told me. When it is completed, it will be able to combine chemicals to synthesize a new drug with limited manual operations. This integrated "reactor" will be a simplified version of Cronin's chemical computer, the prototype of which is a rack with a Pyrex container, which is the difference between a sports car and a trolley.
If you have a design file that outputs the reaction piece on a 3D printer, and you know the chemical program, "it is foolproof," Frei said. "From a chemist's point of view, this is an asset."
Reproducibility is the foundation of all good science, and it is especially important for chemists with Cronin aspirations. "When you publish a paper, it is your moral responsibility to make sure that others can copy it," Cronin said. Reproduction in the laboratory has always been a challenge because the chemistry is handmade. Although strict in principle, procedures are similar to recipes in practice, often passed from professors to students, and rely on undocumented subtleties because they are habitual and actually unconscious. When the brewing starts to bubble or the bubble subsides, do you continue to the next step? Cronin said that this usually depends on expertise: "A lot of what we do in the laboratory cannot be reproduced because our level of expertise is not well explained." But in Cronin's laboratory, the computer Is an expert. Frei's response pieces are reliable, because every step of the automation system must be told programmatically, otherwise they will fail severely.
Cronin has been developing software for the past few years—he calls it a "chempiler"—that can automatically compile every step of every chemical laboratory program, as well as all the equipment and materials needed. Chempiler can extract all of this content from the common language of research papers and mark ambiguities in the paper. After solving all the ambiguities, the chempiler code can run chemicals through the reactor, or the bulky flask and piping system that Cronin showed me when he first introduced me to the chemical computer concept.
Cronin believes that any laboratory can use his free plan to assemble this device for less than $10,000. Although the glassware version is more primitive, he insists on using it out of pragmatism. He hopes to provide researchers with as many choices as possible, and hopes that chempiler will become a "chemistry universal programming language", standardized enough to allow everyone to collaborate. There is a lot of inertia to overcome. "The chemist is grumpy," he said. In order to stimulate interest in the system, he demonstrated its ability by letting a computer synthesize the active ingredients in Viagra.
The computer's ability to manufacture drugs on demand has aroused great interest from many pharmaceutical companies. Although drugs from aspirin to Viagra are mass-produced in factories, the industry sees opportunities for chemical computers to customize small batches of personalized drugs that can treat a variety of diseases from cancer to cystic fibrosis. DARPA also expressed interest and provided funding for Cronin. The agency is excited about the possibility of manufacturing reactors on site, which means that the military can synthesize any medicine or material anywhere by sending digital files to a 3D printer.
From Cronin's point of view, chemical computers will make greater waves in research laboratories. "Most chemists spend 90% of their time studying known chemistry," he said. In order to synthesize the molecules they want to create, they have to go through many preliminary steps, just like a chef prepares the ingredients for soufflé (except that each step may take several weeks and is highly toxic). If a computer can act as a sous chef—preparing any known molecule on demand—chemists can focus on innovation. In other words, they will not be distracted by the heavy work that Cronin endured in Birmingham.
An estimated 200,000 replacement chemists in the liberation of the world are only half of Cronin's vision. The other half is to automate the discovery process. He insisted that chemists will not lose their jobs. On the contrary, research capabilities will be enhanced. "If you have an unlimited number of [chemicals] and an unlimited number of people, you can also conduct an unlimited number of experiments," he said.
His basic idea is to connect a chemical computer with a machine that can analyze chemical substances in real time, add some artificial intelligence and give the system a goal. Then let it run in a closed loop until it hits the bullseye. The goal may be as weird as creating artificial life, or it may be as practical as finding a medicine that treats diseases with minimal side effects.
Other chemists are cautiously optimistic about this vision of automated chemistry. "Lee's work here is very important," said Andrew Cooper, a chemist at the University of Liverpool, one of the pioneers in the field of chemical automation. He was particularly impressed by the scalability of the computer: one day, chemists can seamlessly shift from research to production, mass-produce valuable new materials.
Alexander Godfrey developed early automated systems at drug manufacturer Eli Lilly and now leads the National Institutes of Health's automated drug discovery program. He is more invested in the concept of Cronin. He plans to build his own chemical computer, integrating Cronin's framework and innovation.
Godfrey pointed out that predictions about the future of chemical automation have a tortuous past. In particular, pharmaceutical companies spent a lot of money in the 1990s on systems designed to run multiple experiments in parallel. He joked that this is "garbage in, garbage out". But he thinks this time may be different, mainly because artificial intelligence has matured. Godfrey said that Lee's development will not only "affect drug discovery." It will revolutionize material discovery, from more efficient batteries to more efficient biofuels. "Through democratization, you bring more ideas to the table [and you get] a more diverse group of thinkers."
To discover the origin of life, you can try to build a mobile lazy Susan and let it run 24/7. At least that is the approach taken by Dario Caramelli, a postdoctoral student who works from Frei corner in Cronin's busy laboratory. "[We can conduct] thousands of experiments a day [because the machine] is always performing all the steps," Caramelli explained, noting how the armature above the rotating table can deposit chemicals in one dish while washing the second Plate and dry the third plate. A camera monitors everything that happens inside. If any supply runs out, "the robot will send an email."
This lazy Susan is a variant of a chemical computer designed to explore how random, chaotic, simple chemical substances interact in ways that lead to lifelike complexity—essentially the path that chemistry might take in Darwin’s evolution. survey. "What we are doing is mixing up random recipes, then putting them in petri dishes and videotaping them," Cronin said.
The combination of image recognition software and artificial intelligence can evoke surprises, such as unexpected interactions. (In other settings, the camera is exchanged with more complex instruments, such as a mass spectrometer.) The system usually runs in a closed loop. Significant behaviors can be iterated automatically to achieve higher complexity and more realistic quality.
Unlike chemical computers used to discover new drugs, Cronin did not set specific goals in advance. "I don't know what I'm looking for," he admitted humbly, rare. Since there is no record of the beginning of life on Earth, his goal is to explore as many possibilities as possible without making any assumptions. This chemical method is only feasible with rapid automation.
Walker, an astrobiologist at Arizona State University who worked with Cronin, thinks this approach is attractive because it is unlikely that any life on other planets will follow the same path on Earth. By expanding the scope of the target, understanding life in general can help researchers identify life in an extraterrestrial environment. From Cronin's perspective, creating artificial life-even lifelike behavior-is interesting in itself, because it supports his other grand ambition: the creation of a chemical brain.
Cronin is not particularly interested in the brain. He believes that intelligence is like life, but a chemical phase change. Therefore, in order to induce intelligent behavior in chemicals, Cronin once again bet on random chaos and increase the probability of benefiting him through the speed and efficiency of automation. "I thought, why don't you take a bunch of chemicals and connect them to the electrode array?" He said, and led me into a locked room where he was trying.
The general idea is to expose the gel to an electrical pattern until the chemical substance self-organizes in a way that recognizes the signal, which is the basic form of pattern recognition that controls the behavior of animals and humans. The technique is similar to that used by computer scientists to train some artificial intelligence algorithms, but Cronin uses a material that is as gelatinous as gray matter. Cronin believes that traditional computer simulation is too simple to be very smart, and it makes sense to create intelligence in the way of a real brain: chemistry.
"People used to worry that I did too many different things," Cronin told me when turning on the light of his future brain. "I said,'You don't understand. I actually only do one thing." From the origin of life to artificial intelligence, he was asking a question: "How does a random chemical system become information processing?"
I asked him what information processing means. He said he was talking about all the incredible things that life does, from evolution to advanced decision-making-when you look at the primitive chemicals that make up life, you won't find this phenomenon. He paused for a moment, letting the mystery of metaphysics go deep into it, and then put it another way: What do we ignore in the things that can be combined to build conscious machines around us?
Now he was just waiting for a mess of chemicals to show him.
Jonathon Keats is a San Francisco-based author and the author of "You belong to the universe: Buckminster Fuller and the future."
Keep reading for as low as $1.99!
Moral AI conforms to human judgment in 90% of moral dilemmas
The coming AI identity crisis
According to artificial intelligence, how to paint like Rembrandt
Sign up for our weekly scientific updates.
Subscribe to Discover Magazine to save up to 70% of the cover price.