Watson Gets A Few More IQ Points
September 24, 2013
Watson impressed the world with a computer’s potential when IBM first introduced it in 2009. According to PC World, the powerful processor is about to get a few more IQ points: “IBM”S Watson Could Get Even Smarter With Power8 Chip.” Watson is currently running on the Power7 chip, but number 8 was discussed at a recent conference and is described as being two to three times faster. IBM’s Power line and Intel’s Xeon processor are used in most of the world’s supercomputing. Whenever IBM upgrades its product it uses Watson as a test dummy and the ultimate marketing tool.
Watson is already a powerful tool used in customer service applications, financial applications, and potentially smartphones. The new upgrade will make Watson and other computers that use the Power more than double their processing speed.
IBM has yet to release Power8, but it will soon:
“Recently, IBM announced its OpenPower initiative, where it will license the Power chips and co-develop an ecosystem around the Power architecture with companies like Google, Nvidia, and Mellanox. Up until now, IBM primarily used the Power design in its own servers. This new initiative makes it possible for cloud services and their technology providers to redesign the chips and circuit boards where computing is done, optimizing the interactions of microprocessors, memory, networking, data storage and other components, IBM executives said.”
More is just around the corner for Watson. I have made the joke before and I will continue to do so, I just hope it is not world domination.
Whitney Grace, September 24, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Will Unbundling Work for Enterprise Search
September 23, 2013
I read “Unbundling: AOL, Facebook and LinkedIn” and reviewed the nifty diagram about Craigslist. I ignored the comments about Mr. Newmark, who stood next to me at a Google event, a while back. I am not sure he knows me as anyone other than a person as eager to leave as a couple of other “guests” at the function in Washington, DC. Young people make me nervous, and I wanted to take my Geritol and watch reruns of “I Love Lucy.”
Based on my understanding of Craigslist, I am not sure if the fancy diagram represents unbundling of Mr. Newmark’s idea or just a natural evolution of online opportunities crated by the consumerization of surfing. Craigslist went live in 1995, two years after the Point service, which Chris Kitze and I started with a couple of other clueless but intrepid Internet chance takers.
That makes Craigslist more than a decade old. The service has not changed too much in my opinion. The disruption of various sectors and the growth of the various topics or categories has been a consequence of user behavior, not part of a predestined grand plan. At Point we had zero idea that a large firm would want to buy our service. We were just trying to cope with traffic growth, technical issues, and advertisers who were contacting us.
Nevertheless, the Unbundling article suggests that innovation occurs because big outfits do not seize the many opportunities which their successful services offer up. Here’s the passage I noted:
There are a swarm of services, often mobile first or mobile only, trying to peel off parts of the Craigslist offer, or do things Craigslist should have been doing. AirBnB is only the most obvious. Chris Dixon has a good note about this here, and Andrew Parker produced this great graphic back in 2010.
I understand the viewpoint. However, I think there are a number of factors operating to make it possible to show, as the graphic in the article does, that Craigslist has been a gold mine of ideas.
First, many of the services which take a component of a successful service and elaborate it require bandwidth. A big outfit like AOL, Facebook, or Google for that matter has bandwidth which may be cluttered with noise. The problem becomes “Which opportunity?” which can produce some wild and crazy decisions.
A second factor is personal motivation and capability. The economy motivates some people to look around for ways to make money. Examples which come to mind is the shared ride service disrupting traditional taxi services in some cities. When people need to make money and have a car, an opportunity is perceived. Using an online existing service to build a ride share service is indeed possible. Some innovators may have time and resources to create a purpose-built system free of big company baggage.
Third, customers may be skeptical of new services from big outfits. I have been reading allegations about misbehavior at LinkedIn. I posted a test write up to see what would happen. Do I use LinkedIn for “real” work? Nope. Will I in the near future? Nope. I want to determine for myself if I can “trust” a big outfit. So, a new company offering a specialized service like those on the Unbundling graphic will get a more positive reception than another service from a giant online outfit. I may be in the minority with this approach, but it works for me.
My broad interest in online information access. One of my particular interests in enterprise search. Will the big enterprise search systems be dis-aggregated or “atomized” as the Unbundling article suggests?
This is an interesting question. Enterprise search is not a consumer service. The giant “one size fits all” solutions have been acquired by even larger enterprise solutions providers. Other large scale enterprise search solutions have just gone out of business like Convera, Delphes, Entopia, Siderean Software, and others. Enterprise search has spawned a free and open source solution set as well.
The present enterprise search market is characterized by a frenzy of consolidation, repositioning, and buzzwords. I track the sector on a daily basis, and I am not sure from point to point what strategies and tactics are actually working.
In terms of dis-aggregation, enterprise search has been supporting big solutions like those offered by Dassault (aerospace services!) and Hewlett Packard (yep the PC maker) to specialized solutions offered by Lexmark (printers!) and Xerox (yep, the copier folks).
There are dozens of specialists offering very specific search-related subcomponents which are like a clutch plate on a Lamborghini Urraco. The cost and availability are likely to wreck havoc for the novice’s budget. Some of the vendors of these highly specialized components (Marklogic’s XML technology or Sail Labs’ natural language processing technology) struggle to connect a problem with a solution that makes sense to potential customers. XML fuzzes into databases and databases mean Oracle or Hadoop. NLP blurs into search and search becomes Google. In short, life can be difficult for a provider of an atomized component.
I would assert the following:
- Enterprise search is neither a giant aggregated solution with a couple of big providers nor a sector characterized by on-going disaggregation. Enterprise search is more like one of those weird compounds like ketchup. Sometimes ketchup works like a homogeneous semi-solid. Other times, ketchup is a runny mess. In short, enterprise search is a polystate market. (Technically ketchup is a non-Newtonian fluid.)
- Vendors are never certain what problem their software solves. Take the examples of XML or NLP. What specific value can be delivered to a company looking to cope with Big Data? I am confident that each of the vendors in these two niches can provide me with a webinar that explains the value of their product. But the problem is, “Can an organization get enough value to buy the company’s solution?” And, a more important question, “Are their enough customers to make one or two of these companies generate $200 million or more in profitable revenue today?”
- Customers often do not know what problem they have to solve. Unlike a person who wants a cheap ride to Palo Alto, an organization’s problems are difficult to pin down. More sales means what? Enterprise search vendors talk about customer support and extracting value from information assets. The decision makers at different levels of an organization cannot explain “more sales” one way. Not surprisingly, enterprise search vendors marketing seem to fix a remarkably wide range of problems. Do these systems deliver “more sales”? Usually not in a direct way. Search engine optimization, on the other hand, does deliver measurable results. Maybe SEO will just “take over” enterpriser search at some point.
Let’s assume my assertions are partially correct. Enterprise search is a market space which may be more difficult to crack than seeking inspiration from Craigslist. Perhaps this is one reason why enterprise search has in fifty years produced only one firm which generated more than $800 million in revenues from search technology. Google, please, keep in mind, is in the advertising business. Search is plumbing which helps the company make money.
Conflating findability with the ketchup of enterprise search and its subcomponents delivers one big disconnect in my opinion. Vendors are finding it more and more difficult to demonstrate value delivered by search solutions. Making money with enterprise search is not impossible. I just think it is difficult and getting harder.
Getting inspiration from AOL, Facebook, or whatever big online service one picks may be an easier way to make a buck.
Stephen E Arnold, September 23, 2013
Open Source Community Has 50 New Apps To Brag About
September 23, 2013
We love open source, not just because they offer free software and save us money, but also because the community downright rocks. Here is another reason open source rules from Datamation: “50 Noteworthy New Open Source Apps.” Datamation likes to compile a list of open source apps every once in a while to help its readers be knowledgeable about the latest projects because new ones pop up everyday. When they were making the list they found these interesting trends:
“First was the sheer volume of projects surrounding JavaScript and Web development. Many of these relate to the creation of mobile apps and/or HTML5 technology. The second trend is closely related to the first—cross-platform or platform-independent apps. Web-based apps that will work on any operating system are becoming much more common than those designed for a particular OS. Last is the trend toward open source principles spreading beyond software. Every project on the list includes open source code, but in some cases that code is being used to write an open source book or to do open source science where anyone is welcome to participate.”
Browse through the list and you will find everything from database tools to Web development, which takes up more than half the list. Beyond basic development tools, there are apps for fonts, games, videos, task management, and forums. Some of the apps require a little code savvy, while others can be downloaded with zero to none. We love useful lists here and this is one of the best we have found.
Whitney Grace, September 23, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Google Claims It Is Legally Untouchable Outside US
September 23, 2013
Suddenly, Google’s actions in Europe make a lot more sense. ZDNet reports on a recent claim from the company in, “Google: We Are Above UK Privacy Laws.” This is an interesting development in our increasingly globalized, digital society.
At the heart of the matter is a legal action brought forth by a group of British users of Apple‘s Safari browser, who charge Google with illegally tracking them online. The company doesn’t bother to deny the allegations, but does insist it cannot be prosecuted outside its home country. As for U.S. officials, the FTC already found that Google did, indeed, circumvent Safari privacy settings. The write-up explains:
“The FTC charged Google with placing advertising tracking cookies on computers and devices without authorization. The firm was fined $22.5m by authorities in the United States after the FTC’s ruling. In response to the campaign group’s allegations, the search engine giant says in legal filings that as an American firm, British privacy laws do not apply, and so the matter cannot be brought to a U.K. court. . . . The tech giant does not believe the case is of a serious nature, and according to the claimants, said ‘the browsing habits of internet users are not protected as personal information, even when they potentially concern their physical health or sexuality.'”
To the cynical (or realistic) among us, this stance comes as no surprise. Still, seeing it in print is a little startling, as a few claimants quoted in the article attest. On the heels of Google’s recent “admission” about privacy expectations within Gmail (prompting outrage at many news outlets but given some perspective here), this assertion does nothing to help Google’s reputation on confidentiality.
For those of us who approach the Internet with the idea that nothing that traverses it is truly private, Google’s position on privacy is a shruggable revelation. This stance on international prosecution, though, is intriguing; I am curious to see where it will lead.
Cynthia Murrell, September 23, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Notification Service for Craigslist Users
September 23, 2013
Searching for something specific on Craigslist? (No, I don’t want to know what. That’s your business.) If so, you might be interested in a service we have run across called NotiCraig.com, now in beta. The description specifies:
“NotiCraig.com is a free tool for automated craigslist searches. Instead of wasting all your time searching around on Craigslist, we’ll send you a notification as soon as the item you want becomes available. There is no signup required and we’re a 100% free service!”
The signup form is straightforward, asking questions about what you’re looking for, how much you are prepared to pay, and where you’re located. Naturally, it also asks for your email address so the service can alert you when your item pops up—it will automatically send out an email as soon as its scanning algorithm spots a match. That can be a real help when you’re on the hunt for something that is in high demand.
Cynthia Murrell, September 23, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
A New Acquisition for Yahoo in IQ Engines
September 22, 2013
Our globetrotting leader tells me he had trouble accessing Yahoo in Europe just last week. Nevertheless, acquisitions continue, as the company announces in, “Yahoo Buys Image Search Specialty Startup” at Yahoo UK & Ireland Finance. Service? Well, maybe later.
The startup in question is IQ Engines, and Yahoo plans to incorporate their tech into Flickr (which Yahoo picked up back in 2005). The press release tells us:
“IQ Engines is known for software that analyzes, sorts, and categorizes images using techniques including facial recognition. In May, Yahoo unveiled a dusted-off design of its Flickr photo platform with chief executive Marissa Mayer saying her goal was to make the online service ‘awesome again.’ Sunnyvale, California-based Yahoo has been on a buying spree since Mayer became chief last year with a mission to revive the withering Internet pioneer. The list of more than 20 acquisitions includes Qwiki, a New York operation behind an application that converts video and pictures on iPhones into sharable movie clips complete with music soundtracks. Yahoo in June completed a billion-dollar deal taking over the popular blogging platform Tumblr, a move aimed at bringing more youthful users into the company’s orbit.”
Yes, “more youthful users” are an important commodity for Yahoo, which has indeed been going through some changes since Marissa Mayer took over last year. The CEO’s strategy also includes increased emphases on mobile devices, video services, personalized content, and the company’s reputation overseas. We have one suggestion for Mayer—focus a bit more on service, so Yahoo can retain existing users while pursuing new ones.
Founded in 2007, IQ Engines is based in Berkeley, California. Their image recognition platform, merrily titled Glow, automatically tags and arranges users’ photos and is designed to integrate with both mobile and web applications. The company professes that Glow can recognize people, objects, landmarks, and text within images. We can see why Yahoo wants to add that capability to Flickr.
Cynthia Murrell, September 22, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
When Billionaires Squabble
September 22, 2013
In the Java-related battle between Google and Oracle, a court ruling has not settled the issue. Oracle is appealing the decision by the U.S. District Court of Northern California, which found that the Java APIs Google used to build Android are not copyrightable. Meanwhile, Oracle CEO Larry Ellison seems to have made the issue personal; in a recent CBS interview , Ellison painted the way in which Google CEO Larry Page approached the issue as “evil.” Writer Seth Rosenblatt reports:
“After saying in the public Google+ post published Sunday that Google doesn’t like to get into public battles with other companies, Schmidt then said that ‘Ellison’s claims that Google “took [Oracle’s] stuff”‘ are ‘simply untrue.’
‘”That’s not just my opinion,’ Schmidt wrote, ‘but the judgment of a U.S. District Court.’
“After briefly summarizing the ruling against Oracle, Schmidt placed the public spat in the context of the ongoing clashes over patent reform.
“‘Patents were designed to encourage invention, not stop the development of new ideas and technologies,’ he said.”
I have to agree with Page on the patent /copyright issue. There is a lot of work to be done to bring those systems back into alignment with their original goal—to encourage innovation. (As opposed to stifling it, which seems to be the result more often than not these days when it comes to software.)
I suppose Ellison has a right to his opinion on Page’s moral alignment, but it looks like the industry may be on Google’s side (at least this time). The Electronic Frontier Foundation has submitted to the U.S. Court of Appeals an amicus brief [PDF] outlining their support for the original ruling. At the heart of the matter is the role that the open nature of APIs play in furthering innovation, something we hope the appeals court will not take lightly.
Cynthia Murrell, September 22, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Artificial Intelligence: Does Search Come Up Short?
September 21, 2013
In my view, artificial intelligence continues to capture attention. In actual use—particularly in search and content processing—AI evokes from me, “Aiiiiiiii.”
I read “The Unexpected Places Where Artificial Intelligence Will Emerge.” For investors who have pumped cash into various inventions that understand meaning, the article may surprise them. The future of AI is war, Google, Netflix, Amazon, spam, surveillance, robot space explorers, and financial trading.
The only challenge for AI is its lack of consistency. Smart systems work in certain circumstances and fail miserably in others. In my ISS lectures next week, I profile a number of systems which are alleged to be incredibly smart. The reality is that the systems are often rigged to generate expected outputs. The problem of “you don’t know what you don’t know” plagues the developers of these gee-whiz systems.
Will artificial intelligence improve search? Well, AI makes search easier for those who are happy to accept system outputs. For those who need to dig deeper, AI systems often produce results which do little to provide fine-grained detail or make it easy to identify suspect results.
For a good example of AI in action, look at Google search results when you are logged in. Examine Amazon recommendations closely. Better yet, watch the TV shows and films recommended for you by Netflix.
Stephen E Arnold, September 21, 2013
HP Autonomy Brings Secure Cloud Sharing to the Workplace
September 21, 2013
HP Autonomy aims to bring stronger security to the cloud with Autonomy LinkSite, a solution that integrates WorkSite, the division’s on-site data management solution, with HP’s public cloud-based sharing and collaboration service Flo CM. Market Wired shares the details in, “HP Autonomy Delivers Proven and Secure Enterprise-Grade Alternative to Consumer File Sharing Services.”
It has not taken long for many of us to get used to today’s cloud-based, consumer file-sharing technology. We want to be able to share anything of any size with anyone from any device, synchronizing instantly. Such expectations brought into the workplace from our personal habits can mean real security headaches for businesses. At the same time, continuing to rely on the very limiting method of sharing files through email is becoming less and less tenable. The press release tells us:
“Autonomy LinkSite combines an enterprise-grade document and email management system with the ease of use and simplicity of a consumer solution. It provides the enterprise with a single, integrated, user-friendly tool for external file sharing and collaboration. Autonomy LinkSite enables a single file or an entire project folder to be shared in the cloud with internal and external collaborators, directly from the Autonomy WorkSite application. . . .
“‘For the first time, organizations no longer have to turn a blind eye to continued use of undocumented consumer file sharing services,’ said Neil Araujo, general manager, Enterprise Content Management, HP Autonomy. “Businesses now have a very attractive alternative that satisfies the needs of the users as well as the IT and compliance teams.'”
The write-up lists the following benefits of this new tool: collaboration across firewalls; the convenience of a single point of access for each user; synchronization across all employee devices; an ease of use that they say surpasses that of the consumer-grade file-sharing options; and, perhaps most importantly, the “seamless” extension of security, authorization, and audit properties from WorkSite into the cloud.
Tech giant HP purchased Autonomy in what was, let’s just say, a much-discussed deal back in 2011. Founded in 1996, Autonomy grew from research originally performed at Cambridge University. Their solutions help prominent organizations around the world manage large amounts of data.
Cynthia Murrell, September 21, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Lexmark to Purchase Saperion
September 21, 2013
Will document management work as a new business focus for an old printer company? We should find out soon enough; BusinessLexington reports, “Lexmark Announces Plans to Buy European Software Company Saperion for $72 Million.” The move is expected to expand Lexmark’s data management capabilities; it is probably a wise shift as the transition away from paper records continues. The write-up tells us:
“The purchase is another in a line of acquisitions aimed at transitioning Lexmark from a company that makes printers to a document management company, leaning heavily on software to store and deliver documents without the use of paper except when necessary.
“In addition to having a multilingual structure, Saperion’s platform can integrate with all major business software known as enterprise resource planning (ERP) as well as email and document management systems. Saperion has also developed cloud-based and mobile solutions to provide access to content, even when those trying to access documents are away from their offices.”
Saperion brings along a roster of clients from mid-sized businesses to global enterprises. The new addition would report to another acquired division, Perceptive Software, which Lexmark snapped up last year. I fact, Lexmark has been on a spending spree recently. The Perceptive deal was announced at the same time as the company’s acquisition of Nolij, and this past march it bought up Twistage and AccessVia. It looks like Lexmark is fully committed to adapting to market changes through acquisitions. Will it pay off?
Founded in 1991, Lexmark is headquartered in Lexington, Kentucky. The company stresses that its background in printing uniquely qualifies it to handle unstructured data. Saperion is headquartered in Berlin, Germany, has been helping clients digitize records since 1985. That business likes to focus on simplicity and efficiency, and serves organizations in a range of fields.
Cynthia Murrell, September 21, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
-
- Subscribe to Beyond Search
Feature archive
News archive
-
