The rise of IBM’s supercomputing platform Watson

Last week at the South by Southwest conference in Austin, I attended a networking event and within one hour happened to talk to representatives of 2 startups that use IBM’s Watson supercomputer to enhance their services with smart functionality. Only one day later, IBM and Twitter launched Insights for Twitter, a Watson-powered service for clients to quickly analyze millions of tweets for sentiment and behaviour.

It is a human characteristic to look for patterns in the noise, and here is what I believe to be such a pattern: 2015 is going to be the year of Watson. Now, the issue with the human brain’s pattern recognition is that it works rather bad (unlike the pattern recognition capabilities of Watson, by the way). I could be wrong, at least in regards to the importance of IBM Watson for the general public. However, the evidence is strong that IBM’s artificial intelligence platform actually will play a pretty big role for our immediate digital future and for startups and businesses who want to shape this future.

Watson became somewhat of a media darling in 2011 when the IBM research project managed to win Jeopardy. After the system had proven its cognitive computing capabilities IBM continued the development of Watson and eventually created a new business division based on the technology. The new unit launched last year and was funded by IBM with $1 billion, of which $100 million are to be invested in startups. The newly constructed headquarter of the Watson division, built for more than 600 IBM Watson employees, was opened in October 2014.

So basically, 2015 is the first year in which everything is in place for Watson to really get down to business. Since its Jeopardy victory in 2011, Watson has been made “24 times faster, smarter with a 2,400 percent improvement in performance, and 90 percent smaller”. It has now the size of “three stacked pizza boxes”.

IBM offers Watson’s cognitive computing features in two major ways: One is through direct access for all kinds of professional customers – companies and research facilities who use Watson’s various big data and analytics services to solve their business’ own problems (one of the more highprofile use cases is the effort to fight cancer using Watson). The other way is through Watson’s Developer Cloud, which consists of a bunch of APIs which developers of mobile and desktop apps can use in order to add a bit of Watson flavor to their products. According to the industry media, the Developer Cloud which was launched in October 2013 so far has been integrated into more than 6000 apps. During its current beta period, these developer services are free to use. Here is an overview of the services available.

Some of Watson's services available for developers.

Some of Watson’s services available for developers.

Being neither a developer nor an expert in cognitive computing, machine learning and big data analytics, I cannot say anything about the performance of Watson’s services and how it compares to competitors. But in my role as a curious observer, a couple of things are getting clear to me:

  • IBM has been very smart in first turning Watson into a media darling and then using the prominence to make it a business. Watson is basically a household name.
  • There is talk about artificial intelligence (AI) everywhere nowadays, and advances in computing have brought us to a point in which fiction becomes reality. With Watson, IBM does capture the attention while at the same time providing complex and simple AI services to companies and developers. Instead of having to built their own systems (which would not be feasible for most anyway considering the complexity of the algorithms), they can use Watson. It is the same basic principle like when Internet companies and startups outsource their hardware infrastructure to Amazon Web Service, Microsoft Azure or Google Cloud Platform, or like apps making use of Facebook’s social graph to let users interact and connect to their friends.
  • Speaking about these cloud giants: With Watson, IBM offers something that is missing from the (public) portfolio of the major Internet companies. That gives Big Blue a competitive advantage. It will be interesting to see how Google, Amazon and Microsoft are going to respond to a rising popularity of Watson’s Developer Cloud.

In a Forbes article from 2013, IBM manager Mike Rhodin is quoted saying that “the idea is to make Watson the operating system of the cognitive era”. In layman language that would mean that IBM wants Watson to power as many computing requests as possible in which  computers need to “think” and to understand natural language. That’s ambitious but judging from how far Watson seemed to have come, probably not unrealistic.

One comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.