作者：Guevara
华尔街将再次体验到科技的神奇力量。一款类似苹果Siri的语音助手Warren即将面世。Warren可以即时回答交易员一些复杂的金融问题，能够帮助投资者节省大量查找资料的时间，将研究时间从几天缩短至几分钟。
最近，一群由前Google工程师和管理层组成的交易科技团队研发出了一款虚拟市场助手“Warren”，它有点像苹果的语音助手Siri，只不过是专门面向投资者。Warren可以即时回答交易员一些复杂的金融问题。
Warren的创造者是马萨诸塞州一家名叫Kensho的创业公司。该公司周三表示已经收到了来自Google Ventures、Accel Partners和Devonshire的1000万美元风险投资。
之所以取名叫Warren，是为了继承沃伦巴菲特（Warren Buffett）的精神。该产品正在由一些基金经理和研究团队测试。
目前，投资者们对将类似Siri这样的语音技术用于金融服务非常感兴趣，Warren正是其中的代表。
据FT，Kensho公司表示，Warren可以即时回答100万个问题，能将研究时间从几天缩短至几分钟。到年底时其题库将扩展至1亿个，并且将可以对语音提出的问题作出回应。
Kensho的CEO Daniel Nadler表示：“投资者可以一边看着电视里埃及抗议的新闻，一边问Warren，当中东地区发生抗议时能源价格如何变化。”
Warren可以回答以下类型的问题：“在四级台风登陆美国后房屋开发商、建筑商和水泥公司的股价是怎么变化的？”
Kensho是30岁的Nadler创办的第一家公司，他曾经在美联储做过访问学者。Kensho表示，Warren是基于纳斯达克的OMX云计算平台，将在未来几个月中向不同类型的投资者收取不同的价格。（华尔街见闻）
=====
Former quantitative analyst and Harvard University Ph.D. candidate Daniel Nadler is not your typical startup guy.
“I’m risk-averse,” said Mr. Nadler, who added he’s “not the type to throw on jeans and a T-shirt and launch a startup.”
But, after feeling frustrated with tools available for “quants”–quantitative analysts studying financial markets–and seeing no sign that better technology was forthcoming, he teamed with Google Inc.GOOG -3.13% engineer and friend Pete Kruskall to fill the void.
The pair co-founded Kensho Technologies Inc. last spring to create a Siri-style intelligent assistant for the financial industry and have raised $10 million from Google Ventures, Accel Partners and other top investors to expand their team.
- Kensho Technologies Inc.
- Kensho CEO Daniel Nadler
Affectionately named “Warren,” the software is designed to answer natural language queries for investors much likeIBM’s “Watson” answers health questions posed by doctors. Mr. Nadler says he chose the name Warren because it made him think of an older, wise and patient uncle that knew all the answers.
“Subconsciously, maybe I was thinking about Warren Buffett too–not a bad association,” said Mr. Nadler of the famed investor.
Kensho’s Warrren is still in the early stages. It can answer questions that are grammatically correct, with the object following the subject, on a limited number of topics. The team, which includes about two dozen engineers, is working to account for all the variations in the ways people can ask the same question.
The program can now answer questions like “What happens to the share prices of energy companies when oil trades above $100 a barrel and political unrest has recently occurred in the Middle East?”
In fact, Mr. Nadler and others tussled with that precise question on July 3, 2013. On that day, oil prices topped the psychological threshold of $100 a barrel, Egypt’s military removed President Mohamed Morsi from power and the beginning of “high driving season” officially started.
It convinced Mr. Nadler that Warren was needed.
“Everyone on the street was trying to put their heads together to try and figure out what it meant,” said Mr. Nadler who was working as a visiting scholar at the time for the U.S. Federal Reserve. “You’ve got a financial, a political and a cyclical event. To answer any one of those requires a highly specialized quant to spend hours scrubbing data and coding.”
The team is also in the process of building a massive unstructured geopolitical and natural world event database – an undertaking that has information about hurricanes for example, but not yet about earthquakes.
Mr. Nadler said although the software is still “brittle”–he likened it to Google search in its early days–it has still attracted a diverse and organic following among his target audience.
“We have people with email addresses from Goldman Sachs, Morgan Stanley, Citigroup, Bank of America, JP Morgan… literally every major bulge bracket bank,” said Mr. Nadler, adding that although there are “dozens” of people from each of those major banks signing up to try Warren, none of their employers are yet customers.
The goal now is to continue adding top engineers to the team to improve the natural language search query function and expanding the unstructured database of geopolitical and natural world events.
By the end of 2014, the company expects Warren will be able to answer more than 100 million distinct types of complex financial questions.
Mr. Nadler said he has no immediate plans to monetize, but will charge a monthly or quarterly fee based on the size of the organization once the product evolves and use increases.
“We’ve raised enough money so we can say like Facebook [did], that we won’t care about revenue for two years, but we are a financial company,” Mr. Nadler said. “We are going to be smart about monetization. We’re going to be very Dropbox or Asana in thinking about revenue.”
No comments:
Post a Comment