The December 2022 issue of IEEE Spectrum is here!

Close bar

IBM Watson's Recent Acquisitions Might Make It a Knowledge Machine You Can Actually Use

Big Blue recently picked up a search engine, a Siri-like digital assistant, and an API that can understand text and images

3 min read
IBM Watson's Recent Acquisitions Might Make It a Knowledge Machine You Can Actually Use
Illustration: Randi Klett; Images: iStockphoto

IBM Watson has been buying some interesting companies and technologies lately:

  • Cognea, a company that had developed a “conversational artificial intelligence program” meant to provide more natural interactions than current voice-controlled assistants like Siri and Cortana.
  • AlchemyAPI, a company that provides natural language processing and image recognition as an on-demand service.
  • Blekko, an alternative search engine known (as much as it was) for its classification and filtering of results

Looking at these capabilities all together provides some hints about Watson’s current weaknesses as well as IBM’s long-term plan for the system. All three are customer-facing technologies which provide different ways to interact with Watson’s collection of data and analytics capabilities.

[shortcode ieee-pullquote quote=""IBM has put a lot of different stuff under the Watson umbrella. There are a lot of disparate technologies that are in there."" float="right" expand=1]

In January 2014, IBM set aside US$ 1 billion to turn Watson into an actual business, because it’s generally been better at generating publicity than revenue. Despite all the potential that Watson the-general-purpose-AI might have, it’s still a bit unclear what Watson the business will do. For a large company that is having difficulty commercializing their own internal research, buying up smaller companies that already have users or customers is a common approach (Marissa Mayer’s buying spree at Yahoo comes to mind). 

“IBM is notorious for producing a lot of 80-percent solutions,” says Seth Grimes, an industry analyst and consultant who organizes the Sentiment Analysis Symposium. “The technology is there, but you can't just open up a box, download the software, and have it run. I see Watson in that situation right now.” That’s in contrast with AlchemyAPI, which Grimes describes as a “market tested” platform that’s already had cutomers for years. 

In addition to its web interface, AlchemyAPI also provides software development kits to support text analysis and image recognition in several programming languages. Watson does not have that kind of interface available, says Grimes. “They could build one, but for a company like IBM, it's faster, cheaper, and perhaps more feasible to buy it than to try to productize something out of their research organization,” he adds.

But for Watson to live up to IBM’s hopes, it will also have to be usable for non-developers. Cognea’s technologoy might be a step in that direction. Apple and IBM already had a deal to provide Siri with Watson data and analytics, but it’s clear that digital assistants have plenty of room for improvement. Current question answering systems (including the ubiquitous Google search) don’t have much context or state, says Grimes—each question is answered independently. But IBM aspires for Watson to have an interactive learning experience, so Cognea’s conversational model might help users give Watson better contextualized and refined questions.

Finally, the purchase of a search engine might reignite the idea that Watson will be a Google search competitor. “IBM aspires to assist knowledge workers,” says Grimes, “and Google is arguably in that same field.” After all, Google search results have evolved beyond simply listing links to web pages—the Google Now app on Android devices can provide all sorts of contextual information without ever opening a browser. 

But the two companies have very different ideas about how to make money with such knowledge. As Chuck McMannis, Blekko’s VP of Engineering explained it in his comment on Hacker News:

Blekko's key mission has always been to try to find the needles in this exponentially growing pile of hay. And it is something that the folks at Watson really liked about our technology when we first met at their outreach program to connect with startups. That is what lead to their asking us to join them, and no, they weren't particularly interested in the stuff we had done to provide more topical advertising signals.

The Conversation (0)

Why Functional Programming Should Be the Future of Software Development

It’s hard to learn, but your code will produce fewer nasty surprises

11 min read
Vertical
A plate of spaghetti made from code
Shira Inbar
DarkBlue1

You’d expectthe longest and most costly phase in the lifecycle of a software product to be the initial development of the system, when all those great features are first imagined and then created. In fact, the hardest part comes later, during the maintenance phase. That’s when programmers pay the price for the shortcuts they took during development.

So why did they take shortcuts? Maybe they didn’t realize that they were cutting any corners. Only when their code was deployed and exercised by a lot of users did its hidden flaws come to light. And maybe the developers were rushed. Time-to-market pressures would almost guarantee that their software will contain more bugs than it would otherwise.

Keep Reading ↓Show less
{"imageShortcodeIds":["31996907"]}