Chapter 18 order

order Walk down the hallway of any scientific institute in the world, and you can easily see a portrait of Einstein taped to the wall through the open door of an office: Einstein wrapped in a coat, absent-mindedly Walking on the snow at Princeton University; Einstein staring intently at the camera, with a fountain pen pinned to the collar of his worn-out sweater; Einstein grinning and sticking out his tongue at the world.The founder of the theory of relativity is almost everyone's common scientific hero, a symbol of profound thinking and free creative spirit. In the early 1950s, Einstein was, of course, the hero of a boy named Stuart Kaufman in Sacramento, California. "I admire Einstein very much. No, I can't use the word worship, it should be called love. I love his view of theory as the free creation of human mind, and I love his view of science as the exploration of the secrets of the Creator." Einstein Stein uses OldOne as a metaphor for the creator of the universe.Kaufmann's first encounter with Einstein's ideas in 1954 is still fresh.He was fifteen years old when he read a popular book on the origins of relativity written by Einstein and his collaborator Leopold Infeld. "I was thrilled to be able to understand this book, or thought I could understand it. Einstein's enormous creativity and free-wheeling mind enabled him to create a world in his own head Come on. I remember thinking, it would be wonderful for someone to do this. I remember crying when he died (in 1955). I felt like I had lost an old friend."

Before reading this book, Kaufman had been a good, if not spectacular, A or B student.But after that, his enthusiasm was kindled, not necessarily by science.He didn't feel he had to follow in Einstein's footsteps.But there is no doubt that he felt an Einstein-like desire to penetrate the inner secrets of things. “When you look at a Cubist painting and you see the underlying structure—that’s what I wanted to explore.” In fact, his most immediate interest in it wasn’t scientific at all.As a teenager, Kaufman was passionate about being a playwright, probing the darkness and light in the human soul.His first work, a musical he wrote with his high school English teacher Fred Todd, was "terrible."But he was thrilled to be taken seriously by a real grown-up.Todd was twenty-four years old at the time, and the cooperation with Todd was a crucial step in inspiring Kaufman's intellectual awakening. "Even though it wasn't a very good musical, if I could write a musical with Fred when I was sixteen, why couldn't I?"

So when Stuart Kaufman arrived at Dartmouth in 1957, he was quite a playwright.He even smoked a pipe, because a friend of his told him that if you wanted to be a playwright, you had to know how to smoke a pipe.Of course, he continued to write plays: He teamed up with his freshman roommate and buddy from high school, Mike McGregor, to write three more plays. But Kaufman soon discovered a problem with the script he had written: the characters expressed many arbitrary opinions. "They babble about the meaning of life and what it takes to be a good person. They just talk about it, but they don't act." He began to realize that he was far less interested in the script itself than in the ideas his characters wanted to explore. . "I wanted to find my way to something powerful and magical hidden—though I couldn't say what it was. When I found out that my good friend Dick Green was going to Harvard to study philosophy, I'm very upset. I wish I could be a philosopher too. But of course I can only be a playwright. To give up being a playwright means to give up who I am conceiving for myself."

He recalled that he struggled for a week to think through: "I don't have to be a playwright, I can be a philosopher! So in the next six years, I devoted myself to philosophy with great enthusiasm." research." Of course, he started with ethics.As a playwright, he wanted to understand the question of good and evil.Besides, what else can he learn as a philosopher?But soon he found himself liking something else.His interest shifted to the philosophy of science and the philosophy of thought."To me they seem to be places of depth," he said. What is the science that can be used to discover the nature of the world?What can be used to understand the science of the mind of the world?

Driven by this passion for knowledge, Kaufman graduated third from Dartmouth in 1961 and went on to receive a Marshall Scholarship from Oxford University from 1961 to 1963.As a result, he did not go directly to Oxford. "I had eight months before I had to report to Oxford, so I did the only rational thing: I bought a Volkswagen and went skiing in the Alps in it. I had an Austrian St. The most prestigious address in the East, the Post Hotel. I park my car in the hotel parking lot and use that hotel bathroom all winter." As soon as he arrived in Oxford he found the surroundings very suitable for him.He can still count the three most exciting academic environments in his life, Oxford being the first. "For the first time in my life, I found that people around me were smarter than me. The Americans had a lot of talent there, too. There were Rhodes Scholars, Marshall Scholars. Some of them were already well-known people. Then and One of our group, David Souter of Magdalene, now serves on the Supreme Court. George F. Will (Geofge F Will, a famous American news critic and columnist) and I Used to always go to Indian restaurants to avoid college meals."

A strong curiosity about science and thinking led Kaufman to choose courses in philosophy, psychology and physiology at Oxford.These courses not only include traditional philosophy, but also focus more on contemporary neuroanalysis of the visual system and a broader simulation of the neural connections of the brain.In summary, the course is devoted to the study of the workings of the mind from a scientific perspective.His tutor in psychology was Stuart Sutherland, who would become another influential figure.Sutherland liked to sit behind his desk and toss his students a barrage of thought exercises with questions: "Kaufmann! How does the visual system distinguish between projections onto two adjacent cones of the retina?" of the two dots of light?" Kaufman discovered that he enjoyed confronting these kinds of challenging questions.He found that he had the ability to come up with options on the spot and to give convincing answers. (“Well, the eye isn’t standing still, it’s moving slightly. So, maybe when you stimulate multiple rods and retinas…”) Indeed, he admits that improvising models like this gave him a habit of a habit.He has been improvising models in one way or another ever since.

But, he also had to admit, ironically, it was this ability to improvise models that led him to abandon philosophy for something more practical: medical school. He laughs: "In a way, I don't think I'll ever be a great philosopher. My argument is: I'll never be as smart as Kant. And unless you're as smart as Kant, There's no point in being a philosopher otherwise. So I should go to medical school. That's not a reasoning, you'll notice." Seriously, it was because he was bored with philosophy at the time.He said: "It's not that I don't love philosophy anymore. It's just that I don't believe in a certain frivolity in philosophy. Contemporary philosophers, or at least philosophers in the fifties and sixties, always think that You're testing concepts and their meaning, not the reality of the world. So you can find out if your argument is sound, appropriate, coherent, etc., but you can't find out if you're right. This eventually led me to dissatisfaction.” He hopes to do in-depth exploration of reality, hoping to gain insight into the mysteries of the Creator. "If I had a choice, I would rather be Einstein than Wittgenstein (Ludwig Wittgenstein, a famous Austrian philosopher)."

More importantly, he dared not trust his frivolous weakness. "I've always been very good at conceptual stuff," he says. "At best, it's the deepest part of my personality, a God-given gift. But at worst, it's slick, it's superficial. Because I have this anxiety, I'm so critical of myself. Say, 'Go to medical school. The sons of those curmudgeon women won't let me talk and show my knowledge. Because I have to care for patients, they will force me to know a lot of facts.'" And it is.But somehow, the medical school and the patients didn't change Kaufman's penchant for playing with ideas.In fact, medicine never really had a chance to change Kaufman.Because he had never taken any pre-med courses, he arranged for him to go to Berkeley University in the fall of 1963 for a year of pre-med education before entering the University of California School of Medicine in San Francisco, across the bay.So, it was at Berkeley that he took his initial developmental biology course.

He was strongly blown away by the course. "There's absolutely shocking phenomena here," he said. "It starts with a fertilized egg, and then this thing gradually develops into an orderly new life, and then into a mature life." For some reason, a single fertilized egg can divide and become different nerve cells, Muscle cells and liver cells, and hundreds of different kinds of cells, the process is so precise that nothing goes wrong.The strange thing is not that the tragedies of being born with imperfections are common, but that most newborns are born flawless. "This remains one of the most beautiful mysteries in biology," Kaufman said. "I was completely fascinated by the problem of cell differentiation, and I immediately fell into deep thinking about this problem."

It was a great time to do this research: from 1961 to 1963, Jacob and Monod had just published their series of papers on genetic circuits.This work won them a Nobel Prize later on. (This is exactly what Arthur read 16 years later while lying on a beach in Hawaii.) And Kaufman was quick to read their point.They argue that any cell contains several "regulatory" genes that act like switches, turning other genes on or off. "Their research became a revelation for all biologists. If genes could turn on and off each other, then there would be genetic circuits. In a sense, the genome would be a kind of biochemical computer. It would be the brain of the whole system." This computational behavior, orderly behavior, somehow determines the difference between cells."

But the question is, how are these cellular differences formed? In fact, Kaufman says, most researchers were (and are, for that matter, even now) not too concerned about the problem.They talked about the "developmental program" of the cell, as if the DNA computer were actually executing genetic instructions the way IBM mainframes execute programs written in FORTRAN: step by step.What's more, they seem to believe that these genetic instructions are precisely organized, as completely error-free by the laws of natural selection as any computational code written by anyone.How could it be otherwise?The tiniest error in the genetic program can turn a developing cell cancerous, or can kill it entirely.That's why hundreds of molecular geneticists have long been working in the laboratory to decipher the exact biochemical mechanism by which gene A turns on gene B, and the activities of genes C, D, and E How does it affect the whole switching process.Everything, they think, is in the details. But the more Kaufman pondered the picture, the more the question of how cellular differences loom large.A genome is like a computer, which is great, but it's not an IBM computer at all.He found that in a real cell, many, many regulatory genes can act simultaneously.So, instead of executing instructions incrementally like a human-made computer, a genomic computer executes most, or all, of genetic instructions synchronously and in parallel.If this is the case, he reasoned, it would not matter whether one regulator gene activates another in a precisely defined sequence.It's whether the genome as a whole can settle down and assemble the active genes into a stable, self-coherent shape.Regulatory genes may go through cycles of at most two, three, or four different configurations, but not many, or the cells would scramble all over the place, with genes randomly switching on and off each other, in a state of chaos.Of course, the patterns of active genes in liver cells will be very different from those in muscle cells or brain cells.But maybe, Kaufman thought, that was the point.The fact that a single genome can have many stable forms of behavior may be responsible for the many different cell types that arise during development. People tacitly assume that details are everything.Kaufman was baffled by this.He knew that the details of biomolecules were obviously important.But if the genome had to be assembled and tuned to perfection to function, how did it emerge from random trial and error of evolution?It's like honestly shuffling a deck of cards and getting a hand of spades: It's not impossible, but it's unlikely. "It didn't feel right," he said. "Don't expect God or natural selection to get this far. If we can only explain the biological order in terms of a great deal of detail that would not have occurred in the process of natural selection, if what we see now has experienced If we had struggled so hard, we wouldn't be here today. Just enough chance in space and time didn't make it happen." Things certainly don't stop there.he thinks. "I don't know why, but I want to prove that order is a primordial thing, not something that has been put in and evolved. I'm consciously trying to prove that, in a genetic regulation system, order is innate and unavoidable. Somehow order exists freely in things, it comes into being automatically." If this were the case, he speculated, then this automatic and self-organizing feature of life would run counter to the laws of natural selection.According to Darwin, the precise genetic details of any organism are the product of random evolution and natural selection.But the self-organization of life itself, that is, order, has a deeper and more fundamental meaning.Order arises purely from the network structure, not from the details.In fact, order is the first mystery of the Creator. "I don't know where I got this urge," he said. "Why did Stuart Kaufman happen to come into this world and become interested in the problem of order? The whole thing is a wonderful mystery. One's mind can be fresh and curious about this problem, and can ask these kinds of questions. It’s both strange and surprising to me. It’s just that all my life I’ve had the feeling that all the science I’ve ever done and loved is an effort to unravel this mystery.” Indeed, for a twenty-four-year-old pre-med student, questions about order were like a persistent itch.What, he wondered, would it mean for the genetic order to exist freely?Well, let's take a look at the genetic circuits found in real cells.They've clearly been refined over millions of years of evolution.But another question is, are they really anything special?Among the myriad possible genetic circuits, are they the only ones that give rise to an orderly stable configuration?If so, they are analogues of a hand of spades.It's a miracle that evolution was lucky enough to produce them.Or is a stable network as common as getting a mixed hand of spades, hearts, clubs and diamonds?Because if that's the case, it's a no-brainer that evolution accidentally selected useful genetic circuits.The network in the real cell would be the one that happened to pass through natural selection. The only way to find out, Kaufman argues, is to shuffle the deck, take a set of perfectly typical genetic circuits, and see whether they produce stable configurations. "So I immediately thought, what would happen if you randomly linked thousands of genes together? What would they do?" Now he knew how to think about it.He had studied neural circuits at Oxford, and he knew that real genes would of course be quite complex, but at least Jacob and Monard had taught us that regulatory genes were basically just switches.The essence of a switch is to go back and forth between two states: an active state and an off state.Kaufman likes to think of them as light bulbs (on or off), or as a logical state (true or false).But he felt that whatever image they were thought of, it was this act of turning on and off that formed the essence of regulating genes.All that remains is a matter of interacting gene networks.While Berkeley's free-speaking movement was thriving on campus, Kaufman spent his spare time on the roof of his apartment building in Oakland.He sat there obsessively drawing a network of linked regulatory genes, trying to figure out how they turned each other on and off. Kaufman was really fascinated by the study of gene networks, even after he finished his pre-med program at Berkeley and returned to San Francisco to start full-time medical school.It wasn't that he was bored of medical school, quite the opposite: He found the medical school program very, very difficult.His teacher required him to memorize mountains of textbook knowledge by rote, and to do homework such as the analysis of the physiological structure of the kidney in excruciating pain.But despite this, he still devoted himself to studying medicine.Studying medicine appealed to his inner boy scout: practicing medicine was a good thing to do in any situation, and at the same time taught him exactly what to do, like pitching a tent in a storm. However, Kaufman continued his genetic online game because he could barely control himself. "I was mad to do this weird science of these random gene networks." He got a C in pharmacology. "My notebooks for pharmacology classes are scribbled with diagrams of genetic circuits," he says. At first, he found the genetic circuit very confusing to him.He knew a lot of abstract logic, but almost no mathematics.The computer textbooks he found in the library did little to help him. "At the time, automata theory was well established, and that theory was about networks of logical switches. These books told me how to synthesize a functioning system, or what the general limits to the function of complex automata were. But I was interested in complex systems Where does order come from? No one was thinking about these questions. Of course, as far as I know, no one was thinking about these questions." So he continued to draw his diagrams of random genetic circuits, trying to get a feel for the behavior of these networks model.When he needed mathematics, he invented mathematical formulas to the best of his ability. He soon discovered that if each gene was controlled by many others, so that the genetic network became as densely entangled as a plate of spaghetti, the whole system would violently oscillate into chaos.If you take a light bulb as an analogy, it would be like a giant Las Vegas billboard where the wiring is out of order, and all the lights on it are blinking randomly, and there is no order at all. Kaufman also thought that if each gene was controlled by at most one other gene, and the gene network was very loosely connected, then the behavior of the network would be too simple.It would be like a billboard where most of the light bulbs just turn on and off like mindless nightclub strobe lights.And that's not the order Kaufman imagined.The genetic bulb he wanted was the ability to organize itself into interesting behaviors, like a swaying palm tree or a dancing flamingo.In addition, he knew that very loosely connected networks were unrealistic: Jacobs and Monard had shown that a true gene is usually controlled by several other genes (today, we know that a typical amount is two to ten). So Kaufman took the middle number.Such grid connections are neither very dense nor very sparse.In practice, to make things simpler, he takes a network with only two inputs per gene.He discovered phenomena that implied special meaning.He has long known that densely connected networks can be very sensitive: if you go deep in and adjust the state of any single gene, say, from an on state to an off state, it will trigger an avalanche, causing the network to cascade like a waterfall. Rolling back and forth endlessly.This is why densely connected networks always tend to be chaotic.They can never settle, but in a network with only two inputs, Kaufman found that switching a gene on and off does not cause successively diffuse fluctuations of change.In most cases, the genes that were touched reverted to their original state.In fact, as long as the two patterns of gene activity don't diverge too much, they tend to converge. "Things became easier. I could see the light bulb tend to go on or off," Kaufman said.In other words, a network of two inputs flickers at will like a light, yet is always able to organize itself into a billboard patterned with flamingos or champagne glasses. order!Kaufman worked on it in all his spare time outside of medical school classes.He drew more and more random network diagrams of two inputs in his notebook, analyzing in detail how each network behaved.It is a work that is both fascinating and confusing.The good news about this work is that the two-input network always seems to stabilize fairly quickly.In the best cases, they can cycle through several different states.This is exactly what happens in a stable cell.The bad news about this work is that he didn't know how the two-input model he made had anything to do with real genetic regulatory networks.Whereas real networks in real cells contain tens of thousands of genes, Kaufman's pencil-and-paper networks can no longer hold up to five or six genes.Tracking all possible states and state transitions of a network of seven genes meant filling a 128 by 14 matrix.To do a network of eight genes requires doubling the matrix, and so on. "And the chances of manual error are simply unavoidable," says Kaufman. "I've been staring at my seven-gene network and can't bear the thought of drawing a network of eight genes." "Anyway, by the time I was in my second year of medical school, I couldn't take it any longer. I'd played this game long enough. So I went across the street to the computer center and asked if anyone could replace me. I programmed it. They said, 'Sure. But you have to pay.' So I pulled out my wallet. I'm happy to pay the money." After deciding to let the computer do the work, Kaufman vowed to go all out: He would simulate a network of a hundred genes.He smiles when he thinks about it.Good thing he didn't quite know what he was doing at the time.Let's think about it this way: a single gene can only have two states: on or off.But a network with two genes can have 2×2, or four states: on-on, on-off, off-on, off-off.A network with three genes can have 2x2x2, or 8 states, and so on.In this way, the possible states in a network containing 100 genes are 2 to the 100th power, which is equivalent to 1 million trillion trillion, that is, 1 followed by 30 zeros.That creates endless possibilities, Kaufman said.What's more, in principle, there's no reason why his simulated network shouldn't be able to roam randomly in this space.He made them connect randomly on purpose, and that would mean that his ideas about the cell cycle had no hope of being proven: the computer would have to go through a million quadrillion state transitions before repeating itself.This will be a cell cycle that has been maintained in various states, and this process is boundless beyond imagination."If it takes a computer one ten-thousandth of a second to transition from one state to another, it would take billions of times longer than the history of the universe to run the computer for a million teramicroseconds," Kaufman said. There is no way I could have done this experiment while I was in medical school!" Just paying for the computer was enough to put Kaufman out of business before he graduated from medical school. But fortunately, Kaufman did not do this calculation at the time.With the help of a helpful computer center programmer, he coded his two-input simulated network of 100 genes and handed a stack of punched cards to the front desk with ease.Ten minutes later, the results came out and were typed on wide report paper.The results were just as he expected, showing that the network quickly and stably settled into an orderly state, with most genes locked in an on or off state, and others cycling through several different morphologies.The forms certainly don't look like flamingos or anything recognizable.If the network of a hundred genes were a Las Vegas billboard with a hundred light bulbs, these ordered states would look like vibrating mottled patterns.But they do exist and are very stable. "It's just so exciting!" Kaufman said. "I felt then and now that my discovery was profound. It wasn't something anyone could have intuitively imagined." The two-input network wasn't roaming through a million quadrillion states, but It was a tiny corner of the space that quickly moved into it and stayed. "It settled, wandering, cycling through five, six, seven, or more states, typically about ten states, in an astonishingly high degree of order! I was simply stunned by the results gone." The initial simulations are just the beginning.Kaufman still doesn't understand why sparsely connected networks work so well.But they were so amazing, he felt the results allowed him to look at genes and embryonic development in a completely new light.He used the original method as a model, improved on this basis, and made countless types of simulations.When, he wondered, did this orderly behavior emerge?Why does it appear?At the same time, he also wondered how to test his theory with real data? An obvious corollary of the models he simulated, he thought, was that true genetic networks must be loosely interconnected.Densely connected gene networks seem unable to settle in steady cycles.He doesn't expect real gene networks to all be just two inputs, like his simulated gene networks.Nature has never been so prescriptive.But his computer simulations and all his calculations led him to realize that, in some statistical sense, genetic networks can only be sparsely connected.When you look at the data, the real gene network seems to be as sparse as the simulated one. So far so good.Another test of theory is to look at a particular organism that contains a set of regulatory genes and figure out how many cell types it can produce.Of course, Kaufman was still at the stage of studying the typical behavior of gene networks, and he couldn't say anything special.But he can certainly observe something statistically related to it.He had always had the assumption that a cell type would respond to the steady-state cycle it belonged to, so his simulations got bigger and bigger.He has been tracking to understand how many state loops will occur as the network simulation scale becomes larger and larger.When he got to the point where he simulated a network of 400 to 500 genes, he concluded that the number of cycles was roughly equivalent to the square root of the number of genes in the network.At the same time, he also used his spare time to go to the library of the medical school to read a large number of difficult reference materials, looking for comparative data of real organisms.He struggled to do so, but it worked out: the number of cell types in an organism was indeed roughly equal to the square root of the number of genes in that organism. And so it went. "Hell, I actually made it!" Kaufman said.It was one of the most glorious things that happened to him.By the end of his second year of medical school, his computer bills had racked up hundreds of dollars.But he paid it off without regret. In 1966, at the beginning of his third year of medical school, Kaufman wrote a letter to MIT neurophysiologist Warren McCulloch, explaining his work on gene networks , and asked him if he was interested. Kaufman admits that writing the letter was a bit reckless.Carlo, himself originally an MD, is one of the titans of neurobiology, not to mention his contributions to computer science, artificial intelligence, and the philosophy of mind.Over the past two decades, he and his devoted followers have discovered the hidden meaning of ideas, the results of which were first published in 1943 by him and the eighteen-year-old mathematician Walter Pitts (Walter Pitts). Wrote a thesis entitled "Logical Calculus of Intrinsic Neural Activity".In that paper, Carlo and Beetz claimed that the brain could be simulated as a network of logical operations, such as "and", "or", "not" and so on.At the time, to put it lightly, this was also a revolutionary idea that had a huge impact.Carlo-Pitts' model was not only the first illustration of what is now called a neural network, but also the first attempt to understand brain activity as a form of information processing—a realization that inspired artificial intelligence and The birth of cognitive psychology.Their model begins by showing that very simple networks of logical channels can produce extremely complex computational results.This discovery was soon applied generally to computer theory. But whether Kahlo is a scientific titan or not, Kaufman feels he is the only scientist who can share his work. "Callo is the only one I know who has done a lot of research on neural networks. And it's clear to me that genetic networks and neural networks are basically the same thing," he said. In addition, Kaufman needs a little support from the outside world at this stage.A medical school education has been a mixed blessing for him.In medical school he certainly acquired the "facts" which he sorely needed as a student of philosophy at Oxford, but these facts were unlikely to provide him with deeper results. "I had to do what I was told to do, and it made me very anxious. In medical school, what one has to do is to master the facts, master the diagnosis, absorb the essence of diagnostic wisdom, and then accurately Go through the whole diagnosis process. Although the process of diagnosis brought me pleasure, it lacked the perfection I was looking for. It was not like exploring the mysteries of the Creator.” At the same time, Kaufman's professors were displeased with his search for the beauty of genetic networks. "One of the most meaningful things about going to medical school is the torture of hard labor," Kaufman said.Round-the-clock shifts and endless demands—"The point is to make it clear to you that the patient is paramount. You have to get up at 4:30 in the morning and start doing what you have to do. I don't mind that at all. But there are some faculty members in medical schools who consider themselves the guards of the hospital, and they think that if you don’t have the attitude you need to be a doctor, you’ll never be a doctor.” Kaufman specifically remembers a surgery professor during his freshman year: “He thought my mind was always slipping away. Gave me a D overall. I remember I got a B on my final and he still gave me a D." “所以你可以想象，作为一个医学院的学生，脾气古怪。闷闷不乐，外科得了个D。这对我的情绪有很大的影响。我是马歇尔奖学金的获得者，在学业上一直出类拔萃。而在医学院我却是挣扎度日，我的外科教授告诉我，我是一个多么悲惨的失败者。” 事实上，他在医学院生活的唯一的光明面是他与一个意大利裔美国纽约姑娘伊丽沙白·安·卞奇结了婚。她是艺术系的研究生，考夫曼在牛津遇到她时，她还是个大学生，来欧洲旅行。“我当时正为她撑着一扇拉开的门，心想，嗬，这真是个漂亮的女孩儿。从此我就总是为她撑着门了。”但就连她也怀疑他做的基因网络研究。 “丽沙比我要实际得多了。她对医学兴趣非常大，和我一起去上解剖学课程和其它很多医学课程。但对我的基因网络研究，她的反应是：'挺不错的，但这是真的吗？'对她来说，这网络太虚幻了。” 正是在这种情况之中，考夫曼收到了卡洛的回信：“整个剑桥都为你的研究所激动。”他写到。考夫曼回忆这些时笑了起来。“我一年以后才搞明白，沃伦说这话的意思是，他读了我寄给他的信，认为这很有意思。” 但是当时，卡洛的回信让他又激动又惊讶。他没有想到事情会有这样的结果。他胆壮了起来，回了一封信，解释说，加州大学旧金山分校鼓励医学院三年级的学生走出校门到别处去实习三个月。他问他是否能利用这段时间来麻省理工学院，和卡洛一起做研究？ 卡洛回信说，当然可以。而且考夫曼和丽沙这段时间可以住在他家。 他们立刻就接受了邀请。考夫曼永远也忘不了他第一次见到卡洛的情形：那是在一个冬夜，大约九点钟左右，他和丽沙开着车在黑暗而陌生的马萨诸塞州剑桥街道上转来绕去。他们穿越过整个美国开到这儿，却完全迷了路。“这时他们看见留着长老般胡须的沃伦隐约出现在迷雾中，把我们迎接到他的家里。”沃伦的妻子鲁克端出了奶酪和茶水来款待筋疲力尽的客人，卡洛打电话给麻省理工学院的人工智能小组的第一号人物马文·明斯基（Marvin Minsky）说：“考夫曼来了。” 卡洛是个虔诚的教友派教徒，也是个体贴而又迷人的主人。他高深莫测又奔放不羁，心灵自由地驰跃在广阔的知识天地，以无穷的热情探索思想的内在活动。他行文古风颇健，文章旁征博引，充满了从莎士比亚到圣·伯纳芬图拉（Saint Bonaventura，十三世纪意大利哲学家）的至理名言。然后给文章取名为：《幻想从何而来？》、《思维为何存在于头脑之中？》和《穿越玄学家的洞穴》。他喜欢猜谜、喜欢敏捷巧妙的对话。他是世界上少数能说得过考夫曼的人之一。 考夫曼说：“沃伦常常会把你拖入一个冗长的谈话之中。”曾经住在卡洛家里的学生讲过如何为了避免被卡洛拖入冗长的谈话而从楼上的卧室越窗而逃的故事。卡洛常常会跟在考夫曼后面一起进入浴室，在考夫曼淋浴时，他就放下马桶坐圈，趁着考夫曼忙着把肥皂沫从耳朵里清洗出来时，坐在马桶上愉快地大谈网络及其各种逻辑功能。 然而最重要的是，卡洛成了考夫曼的良师、引路人和朋友，就像对待他的所有学生一样。当他了解到考夫曼来麻省理工学院的目的是要在计算机上做庞大的计算机模拟，从而获取关于网络行为表现的详细统计信息时，他把考夫曼介绍给了明斯基和明斯基的同事西摩·派珀特（SeymourPapert），他们安排考夫曼在当时被称为“MAC计划”的高功能计算机上进行他的模拟。“MAC计划”的意思是机器辅助认知（Machine－Aided Cognition。MAC是取每个词的头一个字母的缩写）。卡洛又安排了一个在计算机编码方面比他懂得多得多的本科生来帮助考夫曼编写程序。最终他们做了上千个基因的计算机模拟。 同时，卡洛还把考夫曼介绍给了虽小但却十分热情的理论生物学界。正是在卡洛的起居室里，考夫曼见到了神经生理学家杰克·考温。杰克从五十年代末至六十年代初在为卡洛当研究助手，现在刚接受恢复芝加哥大学理论生物学小组的委托。正是在卡洛的办公室里，考夫曼见到了英格兰萨塞克斯大学的布朗·哥德文（Brian Goodwin），从此和他成了最亲密的朋友之一。“沃伦就像我的高中老师弗莱德·托德一样。他是第一个认真把我当作一个青年科学家，而不是一个学生来对待的人。”考夫曼说。但令人悲伤的是，卡洛没过几年就去世了，那是在1969年。但考夫曼仍然有点把自己看作是他的事业的继承人。“沃伦一下子就把我带入了那个我从此再也没有离开过的世界。” Indeed.在来麻省理工学院之前考夫曼就决定了，毕业后他要弃医从事科学研究。正是通过卡洛所认识的这群朋友将他真正引入了这个圈子。 他说：“正是由于杰克·考温、布朗·哥德文和其他人，我才于1967年被邀请参加了我生平第一个科学会议。”这是由英国胚胎学家康拉德·沃丁顿（Conrad Waddington）召开的一系列理论生物学会议的第三次会议。“在六十年代中期到后期的那段时间召开的那些会议所做的尝试，正是今天的桑塔费研究所在做的事。”考夫曼说。“真是太好了。从清晨四点起来抽血、化验大便样本——就是我们所谈论的亲手接触现实！——我飞往意大利北部科莫湖畔的色贝劳尼别墅。简直是棒透了。那儿到处是令人惊奇的人。约翰·梅纳德·史密斯（John Maynard Smith）在这里、雷内·托姆（Rene Thorn）刚创立了突变论、芝加哥的狄克·刘文廷（Dick Lewontin）在那儿。狄克·莱文斯（Dick Levins）从芝加哥赶来。刘易斯·沃尔普特（Lewis Wolpert）从伦敦赶来。这些人现在仍然是我的朋友。” “我在会上做了学术报告，介绍基因网络中的秩序、细胞类型数等等。报告结束后，我们走出来，到俯瞰着湖水的阳台上喝咖啡。杰克·考温走过来问我是否愿意来芝加哥做研究。我几乎不假思索地脱口答道：'当然愿意！'足有一年半时间我都没顾上问杰克，我的薪水到底是多少。”