“I really dislike the idea of ​​artificial intelligence. Now the more accurate statement is the automation of decision making, how do we collect data, use data for analysis, and let the machine make as many decisions as possible.†He is in front of the whiteboard. Write down these technical terms and indicate which layer they belong to IaaS, PaaS, and SaaS, and explain to the First Financials the point of origin such as Amazon, Microsoft, and IBM.
AmrAwadallah is the co-founder of Cloudera and the CTO of this machine learning and data analytics platform provider for the IPO in the US this year. In July 2000, Amr's first startup, VivaSmarTIsan, was acquired by Yahoo. He later served as vice president of Yahoo's Intelligent Products Engineering division. In 2008, he founded Cloudera.
"Our goal is to be able to use analytics correctly to solve problems that have not been solved today." Amr positioned himself to do things in the field of artificial intelligence.
Breaking open with open sourceAmr has made an open source architecture, starting with the Hadoop architecture and slowly extending 25 other analysis and resolution tools to form a shared data experience platform.
Building an open source system that provides a shared data experience is what Amr believes is the automation of decision making. “Customers are not technology companies. They want to use technology to solve their own business problems. Except for companies like BAT, companies don’t even have the motivation and ability to do software programming. What they should think about is how to use them. This platform is going to be a better bank, hospital, telecommunications company, etc.," Amr said.
From Amr's point of view, Cloudera is competing with Microsoft and Amazon. On the one hand, the lowest-level infrastructure provided by Microsoft and Amazon, namely IaaS, on the other hand, on the PaaS platform, customers have a need for SQL, machine learning, search and batch processing. In this area, Cloudera has to face the competition with the giants. "Amazon will operate independently by different teams for different needs, but Cloudera hopes to open these services so that users can manage and monitor on a unified platform while meeting government and audit requirements," Amr said.
JP Morgan has developed a suite of software for reviewing their relevant legal documents and contracts using the Cloudera platform. The lawyers have high service fees. Machine learning allows the machine to review most of the contributions. Only very special terms are completed by lawyers. . Amr said that a 10-minute machine audit can save 14,000 hours of work time for humans.
Today, this company that has received Intel investment has a market capitalization of $2 billion.
Amr is more inclined to regard IBM's Watson as a direct opponent. "Watson is a black box, you have to give the data to them, you don't know how they train. The data is power, Watson will provide you with a system that looks like a doctor, but as a customer loses its own value. Values, such as hospitals, manufacturing, you will lose control. You will not be locked up by us, but to machine learning, what we do is visible to you. How to serve people, and our open source process is completely Customers can see that they still retain intellectual property, otherwise it will be a harm to customers, because they will no longer have intellectual property rights, we are not competing with them," he said.
In terms of the new model, if you try to open source and let enterprise customers see the whole process of machine learning, how can the core competitiveness of this enterprise be guaranteed?
In order to maintain the uniqueness of the business model, Cloudera will further add other proprietary business core modules, such as modules for supervision, distribution, security, encryption, auditing, and data backup is not open source.
So why is such software not yet popular? Such software is usually developed jointly with customers, and some people are willing to take the initiative to develop third-party. "The law firm itself may not have the motivation to do such things, because this will infringe on their interests, but our customers are This service is needed. At present, Cloudera's top three industry customers are from the financial, telecommunications and government sectors. The government mainly uses them to study some network security securities and anti-money laundering issues.
This is where business opportunities lie, and why when the extraordinary technology giants bring changes to the penetration of various industries, there are still new companies in the field of artificial intelligence that are emerging and challenge the giant hair.
Chinese companies catch upWhen it comes to competition, China is a city that cannot be avoided.
Amr does not deny this. "China is a very important city." In China, there are also many companies that cooperate with Cloudera. For example, Midea has fully integrated more than 10,000 flagship stores and 300 million users through cooperation with Cloudera. A large number of data from many partners, suppliers and distributors, re-engineering IT systems and achieving centralized management to solve the three core problems of finding users, adapting products to user needs, and completing sales at reasonable prices.
"They (customers) will use our open source software and don't need to give us money until they find their added value. You can use this platform to make constant attempts to let the robot make decisions, once it can really be put into production, Come find us to commercialize this, and come back to us when it comes to maintenance, operations and optimization,†Amr said.
In fact, open source systems follow different protocols. The starting point of the GPL is the open source/free use of code and the reference/modification/derived code for open source/free use, but does not allow modified and derived code to be released and sold as closed source commercial software. There is also a BSD open source protocol, which allows users to freely use, modify source code, and re-release the modified code as open source or proprietary software.
Cloudera chose the BSD open source protocol, which is destined to be a long-term road. According to documents disclosed by Cloudera in the SEC, the company is still in a net loss state despite revenues of $261 million and $169 million in 2016 and the first half of 2017, respectively. "Our goal is to have 8,000 companies with annual revenues of more than $10 billion," Amr said.
At this point, Chinese companies are also playing their own role in the open source system, such as Huawei. "We don't do things like (soft and hard solutions), but benign competition can better promote market development."
As a laborer, how should the application of AI bring about changes in work? This is a question everyone cares about. Amr only expressed his personal opinion: "We may no longer be able to go out as a lawyer after graduating from school, and then work for 60 years. The future work, in addition to really need 20% of experts, the remaining 80% People may not have a job.†So his advice is “to be a very fast learnerâ€.
Puff Bar,Puff Bar Plus,Puff Plus,Puff 1600
Shenzhen Kate Technology Co., Ltd. , https://www.katevape.com