In the big Data age, will computers replace humans in one day?

The computer has enough ability to carry out the activities we thought were exclusively human. 1977, IBM Super Computer Blue (Deep Blue) defeated the world's Western Chess Gary Kasbalov (Garry Kasparov). 2011, the United States famous puzzle answer program "Dangerous Situation" (jeopardy!) in the history of the strongest contestants Ken. Jennings (Ken Jennings) lost to IBM's Watson computer. Today, Google's self-driving car has been on the road in California. While Dale Earnhardt, a well-known race driver, is not Enhard, the guardian of the UK (Guardian) (millions of car owners and taxi drivers around the world) is concerned that self-driving cars "could lead to the next wave of unemployment".

We all expect the computer to do more in the future, so many people suspect that there will be something to be done in 30 years. The trend Prophet hopes the answer is yes. The Luddites, who fought the industrial revolution, were deeply afraid of being replaced by machines and would rather stop building new technology altogether. No one on either side has questioned that computer progress will necessarily replace Labor. However, this premise is not correct, in fact, the computer should be complementary to humans, and will not replace mankind. The most valuable industry in the coming decades will be built by entrepreneurs, and the goal of computer development is to empower people, not to eliminate them.

Humans are very different from computers, humans are conscious, they plan and make decisions when they encounter complex situations, but humans are not good at collating large amounts of data. The computer is exactly the opposite, they are better at effective data processing, but it is difficult to make the basic judgment that everyone feels is easy.

To understand how big the difference is, take Google's computer instead of human project (Computer-for-human substitution projects) for example. Google's supercomputer made headlines in 2012, because after scanning 10 million of YouTube movies, it finally got 75% of the accuracy to identify the cat. It seems like a great thing, but you might think that a 4-year-old can do it perfectly. A cheap laptop can beat the smartest mathematicians in some jobs, but some jobs are for supercomputers with 16,000 central processing units (CPUs), and they can't beat kids. Therefore, people and computers should not compare which one will win, which will lose, they are not the same.

Humans are completely different from machines, which means that working with computers results in much greater results than with others. We don't "trade" with computers, just as we don't trade with livestock or table lights, that's the point: computers are tools, not competitors.

Use the computer to develop a career

The fact that computers and humans complement each other is not just a matter of society as a whole, but also a way to build a great cause. I learned this through PayPal's experience. Over the past 2000 years, we've survived the dotcom bubble and grown rapidly, but faced a major problem: Every month it lost more than 10 million dollars for credit card fraud. Since we have to deal with thousands of transactions per minute, it is impossible to review every transaction because no quality control team has the speed.

So we did what all the team of engineers would do, and we tried to automate the solution. First, the technology chief, Max Ralph, gathers a group of math professionals, studies the transactions of the transfer scam, and then writes our conclusions in software to automatically identify bogus transactions and cancel them immediately. But I soon found out that this was not going to work. After one or two hours, the thieves will react and change the strategy. We're dealing with enemies who can adapt quickly to change, and our software can't react.

While fraudsters can evade the automatic detection algorithm, we find it less easy to cheat human analysts. So Max Revkin and his engineers rewrote it into a software program that could be synthetically detected, and the computer could mark a suspicious transaction on a well-designed user interface before handing over a final judgment. With this integrated system, we lost a profit for the first quarter of 2002, with a net loss of $29.3 million 1 years ago. Because a Russian hacker boasted that we couldn't stop him, we took the system to a Russian name, Igor. The FBI is also asking if we are willing to lend fruit to help detect financial crimes. This made Max Rafe say he was "Sherlock Holmes on the Internet".

This symbiotic relationship between humans and machines allows PayPal to survive, with hundreds of thousands of of small businesses willing to accept payments and grow their online businesses. Even if most people have not seen or heard of it, there is no such thing as a solution without human and machine cooperation.

After 2002 years of selling PayPal, I have also been considering, if the human and computer cooperation can be better than the results of singles alone, that can also be based on the core of the plateau to create what career? The other year, I told Alex Carp, the old classmate of Stanford, and the software engineer Stephen Cohn (Stephen Cohen) put forward the idea of a new venture: we can identify terrorist organizations and financial frauds by using the man-machine composite model of the PayPal security certification system.

We already know that the FBI is interested in this system, so we founded Palantir in 2004 to help people reach important conclusions from different sources. The company is expected to achieve a 1 billion dollar revenue in 2014, and the Palantir is known as the "killer app" by the killer, because it is rumoured to have been a force in the US government's search for laden whereabouts.

We cannot share the details of that operation, but we dare say that the gathering of intelligence by humans alone is not enough to protect people's safety.

The two largest intelligence units in the United States are working in exactly the opposite way, with the CIA (Centralintelligence Agency) dominated by spies who emphasise human superiority, and the State Security Agency, the Agency, is dominated by computer-oriented generals. CIA analysts have to filter a lot of noise, so it's hard to identify the most serious threats. NSA computers can handle a lot of data, but machines cannot determine whether someone is planning a terrorist attack.

Palantir's goal is to move beyond the prejudices of the computer software to analyze information provided by the Government, such as the phone records of leaf-gate militants or bank accounts linked to terrorist activities, to mark suspicious activity for inspection by trained analysts.

In addition to seeking terrorists, engineers using Palantir software have been able to predict the location of insurgents in Afghanistan, prosecute High-profile insider trading cases, sweep the world's largest child pornography syndicates, support Centers for Disease Control and Prevention (Centers for disease Control) to curb the spread of diseases caused by diet and to save hundreds of millions of of dollars a year for commercial banks and governments through sophisticated fraud detection systems.

Advanced software makes this possible, but analysts, prosecutors, scientists and financial experts are more important, and without their active involvement, the software is useless.

Think about what the professionals are doing today? A lawyer must explain the solution to the problem clearly in different ways to the client, the opposing lawyer or the judge respectively; Physicians must have the ability to communicate with patients who are not experts; Good teachers do not have to be professional in the subject We must also understand how to cooperate with students ' interest and learning style, and tailor the teaching method. Computers may perform some tasks, but they cannot be effectively integrated. In law, medicine and education, even if there is no good technology to replace the professionals, these professionals give them more room to play.

Put aside the consciousness of information science

The popular word "big data" also symbolizes the bias that machines can take to replace human beings. Today's enterprises have no more information is not satisfied, mistakenly think that more information can create more value. But "big data" is usually silent information. Computers can find patterns that humans have not noticed, but do not know how to compare the rules of different sources, or use them to explain complex human behavior. Only real-person analysts (or the artificial wisdom that exists only in sci-fi movies) can be found.

We're addicted to big data just because we think technology is amazing. We praise the small things that amazing computers can do alone, but ignore the great achievements that humans and machines can achieve Ching each other, because human participation lowers the magic of machines. Watson computers, deep blue computers and increasingly powerful algorithms are cool, but if the company is asking what the computer can solve, it is not the most valuable business. The most valuable business in the future will ask: How computers can help people solve difficult problems.