How did Google Daniel Jeff Dean become the Internet God of War?

In fact, the content described in the G+ post "Facts about Jeff Dean" is not true. However, it is unusual for someone to go to great pains to set up a website similar to "Facts about Chuck Norris" for him. This is because Jeff Dean is a software engineer, and software engineers are usually not like Chuck Norris, the god of war in martial arts. On the one hand, they are not loners, and software development is essentially a collaborative process. On the other hand, they never beat cowboys with submachine guns like Chuck Norris in this video. Note: Chuck Norris is a karate world champion and an American film actor. He also has a more well-known translation name "Rolex", which comes from the kung fu business card "Raptor Crossing the River". At the beginning of his film career, he played Colt, a karate master, in the martial arts film "The Raptors Crossing the River" directed by Bruce Lee, and confronted Bruce Lee at the Colosseum in Rome, which was recognized as a classic martial arts scene. Jeff Dean However, on April Fool's Day in 2007, some young Google engineers thought that they should make a website for Jeff Dean to praise his achievements in programming. For example, the compiler will not warn Jeff Dean, but Jeff Dean will warn the compiler. Jeff Dean writes binary code directly, and then he documents the source code for other developers. When Jeff Dean thinks about ergonomics, it is to protect his keyboard. One day Jeff Dean was forced to invent an API for asynchronous calls while optimizing a function. In this way, this function can return the result before being called. The truth about Jeff Dean is that you must be a computer expert to understand many jokes people tell about Jeff Dean. For interested readers, the Business Insider website provides some explanations about his popular jokes. If you don't have the background knowledge of computer science, it's hard for you to understand the false achievements he mentioned in those jokes, let alone his real achievements at work. Dean's own system, such as MapReduce, BigTable, Spanner, etc. Many Google users don't know that it comes from Google. However, these programs are the cornerstone of the existence of Google and the modern Internet. Some of the projects he is engaged in now may revolutionize information technology again. When you think about who created today's Internet, you may think of founders and CEOs of many companies, such as Tim Berners Lee, mark anderson, Larry Page and sergey brin, and maybe mark zuckerberg. This makes sense. These people have invented a product or framework that shapes the way we use the Internet today. At the same time, in the shadow of these giants freed from heavy daily work, there are a group of unknown developers who knock out products and systems on the keyboard for us to use every day. Unlike other industries, in high-tech industries, these people are usually irreplaceable. A good accountant can save you 5% personal income tax. The basic attendance rate of an excellent baseball player is only a little higher than that of ordinary players. But a good software developer's work in a week may take a team of nearly 10 people several months to complete. This difference is exponential. This is not the fact about Jeff Dean, but the common sense of high-tech industry in Silicon Valley, which is why the best companies spend so much money to attract top talents. When Dean joined Google in the middle of 1999, he was lucky enough to be the top young computer scientist in America. When home computers first became popular, Dean said that he was always looking for ways to pursue the highest performance on a given machine. When he was in high school, he wrote a software to analyze a large number of epidemic data. According to him, his software was 26 times faster than the professional software at that time. This system, called Epi Info, was adopted by the Center for Disease Control and translated into 13 languages. When he was a Ph.D. student in computer science, he studied compilers. Compilers are used to translate program source code into computer-executable languages. He said, "I have always liked fast code." But Dean was not satisfied with the status quo. He didn't want to spend his whole life on compilers, so he left academia later. Less than three years later, he joined Google with only 20 people at that time. (According to Steven Levy's book the Plex, Google, as a search startup, thinks Dean is a hard-won talent. He made great contributions to early Google News and AdSense, among which AdSense, an advertising product, rewritten the rules of the game of Internet companies. After that, he turned to one of the core issues of the company: scalability. The original idea of Google's basic algorithm came from Page and Brin, both top developers at that time. In the late 1990s, they created PageRank algorithm, which is an algorithm that returns the most relevant search results when users give a search query. Concerned about the relevance of search results, Google surpassed Yahoo, AltaVista and other leading search engines at that time. However, as Google became more and more successful, it also encountered huge technical challenges. Dean recalled, "We couldn't deploy more machines fast enough to meet the demand". So Dean and his colleagues, including another excellent programmer, Sanjay Gemawat, found a solution. This problem looks like a hardware problem, just like he treated Epi Info in high school. Gemawater helped lead a team to develop Google File System (GFS), which enables very large files to be distributed and stored on many cheap servers. Then Dean and Ghemawat developed a programming tool called MapReduce to help developers effectively use these machines to process huge data sets in parallel. Just as compilers help programmers write programs regardless of how the CPU handles them, MapReduce enables Google developers to adjust search algorithms or add new functions without worrying about parallelizing these operations or handling hardware failures. Dean and Gemawat's method is so powerful that when they published a research paper at a conference in 2004, it immediately became the industry standard. Until today, MapReduce has become an important cornerstone of many other projects, one of which is the famous open source framework Hadoop. It is Hadoop that has created a new industry buzzword "Big Data". From online travel to energy exploration, big data methods are used in different fields. Just as Google began to expand from MapReduce to other new programming models in some core applications, Dean said that he still saw many summer interns start to use MapReduce extensively when they started new projects after arriving at Google. The 10 times effect mentioned by Page, one of the founders of Google, MapReduce is a good example. The effect of 10 times is better than the original 10 times, not better than the original 10%. MapReduce doesn't make a certain kind of operation faster, but it helps every developer in Google to do things they couldn't do before. Dean's other projects have similar exponential effects. On the basis of Google file system, he and Gemawat created a distributed database system called BigTable. BigTable can handle the data of 1PB (1Pb =100000 GB. (Translator: The open source community has a similar project, based on Hive on Hadoop) Then they further developed the Spanner system, which is called the largest single database in the world. Case Metz of Wired magazine said that by using innovative time synchronization methods, Spanner's physical storage spans different data centers around the world, but it operates as if it were in one place. In other words, it can make different information in data centers around the world consistent, even though a specific update request may take different time to reach different data centers. Metz added that before Spanner was reported, no one ever thought that such a system could be built. Now, these real facts about Jeff Dean seem a little fake. The dean himself will laugh at this situation, saying that it is a bit embarrassing, but it is also a feeling of being praised. But he also said that it should be remembered that his real achievements in these jobs are always achieved through cooperation with different people. Almost every morning, he goes to work at Google headquarters in Mountain View, California, and always sits down to have coffee with the same group of people. He estimated that over the years, we probably eliminated 20,000 cups of cappuccino together. These people don't always work together. In fact, some people have moved to different offices on the other side of the Google campus. But when they get together to discuss what they are doing, some people's questions can always inspire others' new ideas. These coffee chats enabled Dean to apply his experience in optimization, parallel computing and software architecture to many different types of projects. All these have given him enough ambition and confidence. As his long-term partner, Gemawat said, "He is always full of enthusiasm and optimism about what we can do, and nothing can stop him". His recent work can well explain what Google will do next. Last year, together with Andrew Ng, an expert in machine learning at Stanford University and one of the founders of Coursera, he helped Quole, a graduate student of NG, to conduct an unprecedented unsupervised machine learning experiment. This experiment belongs to the secret Google X Skunk project of Google, and uses 16000 processors to learn YouTube videos unattended to find out how to identify a cat. It seems that many computers are used to get a very basic result, but this experiment can help us lay the foundation for the next generation of artificial intelligence technology. Future artificial intelligence technology will play a role in many potential applications, including Google Now, which uses personal assistant technology, and image search function, which is very helpful to Project Glass. Jeff Dean may be inventing something incredible, just like the special keyboard with only 0 and 1 mentioned in Facts about Jeff Dean (translator: this keyboard doesn't exist, and the article on this webpage joked that Jeff Dean wrote the program directly in binary machine code. Jeff Dean admits that he is not an expert in machine learning, but he is willing to use his experience in building scalable and highly available systems to help this research. Contrary to the "facts about Jeff Dean", Dean said that in many cases, the best way to solve the problem is not simply to sit down and start writing programs. His method always needs some simple calculations at the beginning in order to find the best balance between quality and speed for a specific process. He said that in many fields, from machine translation to search quality, you always try to weigh the amount of calculation you can do for each query. Maybe you can't find the ideal solution, but we can always get a 98% return by calculating 1% in some approximate way. Dean often does this kind of calculation, so that he gives a "list of numbers that every computer engineer should know". For example, this includes how many milliseconds it takes to send a network packet from California to Amsterdam at the speed of light (150ms). By remembering these numbers, you can tell which of the three designs will be the best on the whiteboard in 20 minutes. He added, what if you can't calculate quickly? Convert all these numbers to a power of about 2, which will make multiplication easier. If Dean really has superhuman strength, then this ability can't make things perfect in an instant. This is an ability to weigh, optimize and deal with problems according to different levels of things. In other words, it is the ability to find opportunities and do things as well as possible in a short time, rather than trying to pursue perfection from the beginning. In Silicon Valley, this is much cooler than shooting cowboys with submachine guns.