Foreword-Industrial Spirit

The industrial revolution was neither a cotton age nor a age of steam, but an era of progress.

From 65438 to 60' s, the industrial revolution first started from the cotton textile industry in Britain, and then spread to mining, iron smelting, transportation and other industries. From shuttle and Jenny spinning machine to steam engine and internal combustion engine, countless scientific and technological inventions condensed human wisdom have opened this mighty five modern industrial revolutions.

Watt, the inventor of the steam engine, once wrote an autobiography in the third person. In his autobiography, he wrote: "His mind is haunted by how to make a cheap and excellent engine." Besides function and beauty, this pursuit of economic value represents the development peak of European technical rationality in the past 1000 years.

In the era when technology is rapidly leading social change, countless pioneers are attracted by some kind of "inducing factor", which is not only a technological breakthrough, but also a low price. This "inducing factor" caused by economic value symbolizes a new potential, which can ignite the imagination of these shrewd and determined technical and commercial pioneers. In other words, the inducement clearly shows that enterprises based on related innovations will be competitive in cost [1].

Financial and technological innovation

The technological revolution breaks out every forty or sixty years, and the process of change it brings affects all aspects of society. In every technological revolution, financial capital is the most willing and daring customer of products and services brought by the new technological revolution. It is always ready to speed up transactions, expand business areas, and at the same time promote every technological revolution in an indirect but extremely important way.

Every innovation in infrastructure, technology and organization has accelerated the transportation of goods and information transmission. These innovations can usually bring about changes in the monetary, banking and financial sectors. After the opening of the Suez Canal, ships, international telegraph lines and other things that are conducive to international trade are widely funded by capital. In several information revolutions initiated by the United States, banks were early customers of cheap post offices, national railways and telegrams, and were also the first institutions to use telephones, typewriters and calculators.

The Chicago Board of Trade was founded in 1848, the year when the telegraph line was set up in this city. In the next few years, merchants set up similar futures institutions in several other major commercial centers.

1887, Green, then president of Western Union, the largest telegraph company in the United States, said that at least 87% of the telegrams transmitted by Western Union were business-related, and most commercial telegrams were speculative. This kind of speculative telegraph exchange is mostly "no transportation, often no delivery". Therefore, Green firmly believes that telegraph "is essentially an accessory of business and speculation, which needs immediate communication and reply" and is not a means of mass communication [2].

At that time, a business editor once commented that before the commercialization of telegraph, businessmen could speculate on large-scale agricultural products by using distant price information known in advance. After the commercialization of telegraph, because the price data can be transmitted quickly by telegraph, the original speculative arbitrage mode of businessmen no longer exists. Instead, we began to speculate by guessing the possible price of agricultural products on a specific date in the future. In this way, time replaces space as the biggest unknown. According to the editor's estimate, the growers of these agricultural products lose about $40 million a year because of this speculation, but after the commercialization of telegrams, this loss is reduced to one twentieth of the original.

The essence of finance is the transmission of information, which produces expectations of actions and results. The financial system, whether banks, primary market or secondary market, is an information transmission network. Through the information network, when this expectation is transmitted more accurately and quickly, wealth is created. Therefore, every information revolution: language, writing, printing, telegraph, telephone, internet, mobile phone … has created a financial revolution.

And we are now in an era when a new financial revolution is about to happen. The inducing factor this time is artificial intelligence. Artificial intelligence will reshape the financial information network, gradually assist the information transmission and processing channels traditionally carried by experience and contact with machines, and make them automatic and intelligent.

Want information industry instead of information agriculture and mining.

In the field of financial informatization, we have seen countless scientific and technological enterprises pouring in, but all kinds of enterprises, large and small, have also been saved. We will examine these enterprises from the perspective of processing financial data into financial information.

An enterprise model can be called "information agriculture". They rely on manpower to mine data from some "natural resources" and lack the ability to expand reproduction on a large scale. This kind of "agriculture" is difficult to do deep processing of data, with limited value-added data, short industrial chain and limited information value of output.

The other mode is "information mining", which uses some mechanical tools to mine some existing structured data and assemble some more valuable information in a certain scene. This kind of mining is better than agriculture, but the disadvantage is that the categories of mining are limited, many people can do mining, the technical barriers are not high enough, and the value added is limited.

What is the "information industry"? We all know that Ford has rung the bell of modern industrial mass production mode. What makes mass production possible in the historical process? Let's go back to the time when Ford improved the Model T in the early 20th century.

1908, on the eve of the launch of Model T, the average working cycle of each assembler in Ford Company, that is, the working time experienced before starting to repeat the same operation, was 5 14 minutes in total. In the spring of 19 13, Ford took a new step in the new factory building in Detroit Highland Park, which was to install a mobile assembly line. Workers stand in one place without walking around, but the assembly line delivers cars to them. This innovation shortens the working cycle of workers from 5 14 minutes to 1.9 minutes, and the production efficiency of automobiles is improved tenfold and hundredfold. This kind of mobile assembly line is commonly called "assembly line" in modern factories.

Ford was able to innovate the large-scale intensive production mode because it technically overcame the problem of "warping and deformation of mechanical parts after heat treatment", which was the bane of parts that could not be standardized in the past, and always had to be polished manually by mechanics over and over again. Once the warping problem of parts is solved, the number of parts can be reduced and they can be easily connected with each other.

Then, to make the financial information industry possible, we need a technology that can overcome all kinds of "financial data standardization". We need to do deep processing of data and make the analysis of paragraphs, sentences and entities to the extreme. We need to structure and materialize various financial documents and extract their metadata, so as to produce the reorganization and automation of tens of millions of kinds of data. From data to information, and finally to build a deep financial knowledge network, this is the financial information industry we want.

AlphaSense—— A New Generation of Financial Knowledge Engine

On the 20 16 Forbes list of the top 50 American financial technology companies, a listed company named AlphaSense appeared.

Through the analysis of AlphaSense and its competing products, compared with financial information data platform or financial information engine, these products are more like the next generation financial knowledge engine system. But before introducing them, let's review the concepts and progressive relationships of data, information and knowledge.

Data is a signal reflecting the motion state of objective things, which is perceived by sensory organs or observation instruments and forms data in the form of words, numbers, facts or images.

Information is the processing of data, which makes the data related to each other, forms a text to answer specific questions, and is interpreted as meaningful information in the form of numbers, facts and images.

Knowledge is not a simple accumulation of data and information, but information that can be used to guide practice. Knowledge is the synthesis of knowledge and experience gained by people in the practice of transforming the world.

In the scene of investment research, analysts usually need to get a lot of data, information and knowledge from news, financial reports and research reports, and then reorganize these materials into investment decisions through their own logic and world outlook. According to the difficulty of obtaining materials, we are divided into five grades from simple to difficult:

1. Company, industry, market and other fresh information and data (company, stock price, turnover, etc. )

2. Relevant indicators and data information (CPI, freight volume, various charts of industry scale, etc. )

3. Fresh judgments and conclusions (various bearish conclusions)

4. New evidence (factual basis supporting the claim)

5. Other people's logic and research framework (whole network knowledge network)

For those financial information companies engaged in information industry and mining, their products are nothing more than financial data terminals or information platforms. The first three levels of data information can be used by users, but they are weak in dealing with the depth information and knowledge of the last two levels. So how do companies like AlphaSense solve problems for analysts at this level in the organization through products?

Their products consist of three parts: advanced semantic search engine, interactive knowledge management system and document (knowledge) collaboration system. Investment researchers obtain all kinds of "materials" through advanced semantic search engines; In the interactive knowledge management system, data can be collected and managed selectively; In the knowledge collaboration system, materials can be processed and reorganized.

When the knowledge slice becomes fine enough for people to search, manage and reorganize, it is a budding stage of financial information industrialization, and to some extent it is a preliminary financial knowledge network.

Kensho-provoke a new round of military competition

Let's take a look at Kensho, another hot Fintech company. Its founder is a doctoral student at Harvard University, and its engineering team is top engineers brought from Google and Apple, invested by Google and Goldman Sachs. It is said that their artificial intelligence technology makes everyone on Wall Street feel insecure, and their company itself is quite topical. Thanks to various media hype, we always mention it with a little awe. Is it true that machines will replace humans? Even replace those Wall Street analysts who stand at the peak of human intelligence.

Because of the channel, Kensho's products are still a "black box" for most of us and can only be carefully pondered through various peripheral information. Daniel Nadler, founder and CEO of Kensho, is a typical elite. He holds a doctorate in economics from Harvard University and worked as a visiting scholar at the Federal Reserve for some time.

At the annual financial technology conference held by MIT on 20 15, Daniel told the public his whole story as a Kensho [3]:

"When I was at the Federal Reserve, there were a bunch of tools like Bloomberg, Reuters and Capital IQ on my desktop, but according to my insight into politics, weather phenomena, geographical environment and other information, the software I used still couldn't solve the problem of what I wanted to buy, especially some event-driven data analysis ... Later, I went to the west coast to find Google, whose vision was to organize all the information in the world, but why couldn't Google organize me as a financial scholar?

"... I stayed in Google for a few months and told them that since your goal is to organize information all over the world, since there is so much unstructured and organized information in the financial field, shouldn't someone do something? I persuaded Google, so they became our initial supporters and gave us many excellent engineers. "

Daniel Nadler bluntly criticized Mike Bloomberg (the founder of Bloomberg Terminal) for his old-school business, saying that he was like a general who knew how to win World War II (not the future world war).

But there are also some criticisms of Kensho himself. On Zhihu, a user from new york shared his feelings about using Kensho products, and he made a summary: "Bloomberg can easily copy this gadget and make it better 100 times". Let's really sort out what Kensho is doing [4].

Daniel introduced in an interview that Kensho consists of two business lines of two departments. One is the risk analysis department, which helps large banks and other financial institutions understand non-trading risks. Non-trading risks are not market risks, but use historical data to help them analyze the risk exposure caused by factors such as geography and weather. Another big business line is the analysis of global commercial media, and Kensho rebuilt the analysis engine of commercial TV media. Therefore, kensho is a media company on the one hand and a bank risk control service organization on the other.

Let's take a look at the business line of bank risk control first. As a financial expert who once worked in the Federal Reserve, Daniel has the ability and channels to do business. As the strategic investment target of American consumer news and business channel (global financial cable satellite news station owned by NBC Universal Group), the smooth development of media business is also expected. The media nature of commercial TV makes American consumer news and commercial channels have low expectations for Kensho's analytical ability. After all, the most important thing for a TV media is the speed of news. After all, the time a news stays in the audience's mind is very limited. After the incident, give an effective analysis conclusion as soon as possible, which is the key to defeat the media peers. So we can see that Ask-Kensho can bring a stunt to the media on the one hand, and it is indeed a means to improve competitiveness on the other.

From the above analysis, we can see that Kensho's two business lines are both competent and valuable to it, so what caused the rumor that "Kensho replaced Wall Street analysts"? The "culprit" should be Goldman Sachs, because Goldman Sachs is not only an investor of Kensho, but also their customer. Let's take a look at the following report [5]:

Kensho's main customer in Goldman Sachs is the sales staff in the trading hall of the bank. In recent months, they have used software to answer inquiries about buying and selling energy stocks and commodities. These people want to know how they should combine their investments to deal with the raging fire of jihad in Syria.

In the past, these salespeople would summarize according to their own understanding of recent events and market reactions, but they were limited by people's memory. For particularly valuable customers, the sales representative may ask the research analyst of Goldman Sachs to conduct a more complete research, dig up past news events and find out the market's reaction to each situation. The problem with this method is that when the research results come out, the trading opportunities have already slipped away.

Kensho only serves the sales department of Goldman Sachs, helping them to respond to customers' inquiries quickly, so that these salespeople can better sort out shallow information. Another function of Kensho is to help some researchers complete some elementary work, instead of the analyst task of "I used to spend two days a week doing this kind of thing" or "I used to hire people to do nothing but do one thing".

Kensho can indeed serve top investment banks, but it may be just a "beautiful misunderstanding" or a "beautiful wish" to really replace "analysts" and make more people unemployed in investment banks.

However, if Kensho really has the analytical ability comparable to that of Wall Street analysts and becomes a tool available to mass investors, it will actually promote market effectiveness. For example, a few institutional funds observe a signal and use it to establish a profit strategy. However, with Kensho's powerful analytical ability, it will also capture information and make it public, and the advantages of top institutions due to information asymmetry will disappear completely. As more and more market participants use the same strategy, the strategy will eventually fail. Therefore, the market price can reflect "all available information" more quickly and to a greater extent.

Just like the telegraph of19th century, it didn't completely put an end to businessmen's speculation on agricultural products, but it changed the nature of speculation and reduced the losses of farmers who were in a weak position at that time. If Kensho really applies AI technology to a critical state, thus changing the nature of modern financial speculation again, then Bloomberg, Reuters and even other fund institutions themselves will be forced to join the arms upgrade race to avoid becoming laggards from leaders.

summary

In the era of artificial intelligence, AlphaSense is building a financial knowledge engine, Kensho is provoking a new round of arms race, and these pioneers are launching an industrial revolution of financial information. After the prelude of the revolution blows, can we predict whose future is the brightest?

We should also see that just as the first industrial revolution is a long historical process, the industrial revolution of financial informationization will not be completed in just a few years. At this early stage, I am afraid it is still in the stage of building various parts of the "smart financial core engine". Financial analysis is gradually moving towards the standardization of "parts" from "agriculture" relying on manual work and "handicraft industry" relying on experience. The most important thing is not the final vision, but how to design a reasonable development path driven by the market. AlphaSense is based on the discovery and reorganization of financial fragmentation knowledge, and Kensho is based on the rapid dissemination and discovery of shallow information, which are the necessary intermediate links of the ultimate vision. From the prelude to the climax, we should seek truth from facts and don't aim too high.

* Note: Some opinions and information in this article come from the following references *

[1] The lever of wealth: technological innovation and economic progress

[2] "Information has changed America"

[3] The speech of the founder of Kensho uploaded by internet users on YouTube.

[4] Zhihu asked: "Why is Kensho favored by Goldman Sachs?"

[5] Alpha Workshop: "How artificial intelligence company Kensho changed Wall Street"