The future of search only cost $30 million

March 27, 2013

On the surface, Nick D’Aloisio’s story is the kind tech lives for, and sometimes regrets. It’s the tale of a kid selling an obscure startup for an inflated price, and then it becomes as irrelevant as Netscape, and its buyer’s remorse is part of the company’s enduring legacy.

But the story of Summly, a startup whose app appeared in the Apple Store only five months ago and was purchased on Monday by Yahoo for a reported $30 million, isn’t part of this trite arc.

This isn’t a boilerplate tale about a youngster hitting the jackpot, a former Internet giant trying to buy a relevance makeover, or even about an intriguing programmatic way to summarize news. It is about the future of search. 

D’Aloisio’s youth – he’s 17 – and windfall are interesting data points, even if all the work behind the magic algorithm isn’t the sole product of this high schooler’s brain. Like all really good ideas, Summly’s is simple: Anything can be summarized, but by having a computer do it,  the number of things you can summarize — and the speed with which it can be done — are massively increased. As an app, it filtered news stories and — Presto Chango! — spit out the CliffsNotes version, optimized for a smartphone’s tiny screen (and our infinitesimal attention span).

If nothing else, D’Aloisio put together a company with serious backers — the first was when he was 15, and then some eyebrow-raising names like Yoko Ono and Stephen Fry and Ashton Kutcher followed. These investors were either captivated by this young man or captivated by his idea, despite him.

Summly got off to a decent start in downloads, but it was a free app without an obvious business model other than charging publishers to be featured on the app, or selling ads, neither of which are exactly a slam-dunk.

And that’s because, however cool it sounds to have an automated way of summarizing the news, it really isn’t all that exciting. In defense of my profession, that’s pretty much what the  first 2-3 paragraphs are meant to do, no matter how long we ramble on. Taking long stories and making them shorter because smartphone screens are small sounds like the cover story.

The real value in Summly: Not summarizing the news. Summarizing the Internet.

Here’s my wild speculation: Summarizing news stories was Summly’s proof of concept. It was never meant to be a consumer item but was field tested with its true purpose concealed: As a potentially game-changing search engine layer that could dramatically improve the relevancy and coherence of results.

View Summly as part of something, The Semantic Web, just taking form. The Semantic Web is one in which we can “ask” questions any way we want, and know that we’ll be understood. We’d communicate with the Web the way we communicate with each other, without any special rules about syntax or grammar apart from the common language we share.

Siri, Apple’s celebrated iPhone personal assistant, is part of this nascent trend to “humanize” our relationship with computers and data — I likened it to a poor man’s Watson in my tech predictions for this year. Siri’s a good listener — I use it to transcribe all the time — but it’s not very intelligent even though it works with some excellent databases like Wolfram Alpha, IMDB and Yelp, and defers tp Google search when all else fails.

Techonomy.com’s Kevin Haney sees this too:

Summly is to Watson what a Hot Wheels toy car is to a real Indy race car. But still, there’s a link. The kind of thing Summly can do on a cell phone over the network will get more and more powerful—more and more Watson-like. The kind of thing Watson can do will, similarly, get packaged in ever smaller, cheaper, easier services, until it’s available on a cell phone over the network.

Imagine a world where something like Siri interpreted your command, Summly ran across the Internet collecting information, came back in fractions of a second and gave you a perfect reply? No more hits based on a single word in the comments. Content farms could more easily be weeded out. SEO as we know it might be rendered obsolete.

We’d be closer to the data. The methods by which we access the medium wouldn’t matter nearly as much. It would be like living on the bridge of the Starship Enterprise.

What price would you put on that?

Search is a golden plank in the Internet platform, and refinements are always being made by the big three — Bing, Yahoo and especially market-leader Google, whose “Panda” algorithm tries to keep up with the constant need to refine results.

I’m betting that Summly itself develops into a powerful weapon to help search results stay relevant. If that happens, we’ll be laughing at what a bargain it really was.

PHOTO: Nick D’Aloisio, aged 17, who developed the smartphone news app Summly, poses for a photograph at offices in central London March 26, 2013. REUTERS/Suzanne Plunkett

9 comments

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

One still has to question what Y! Got for $30M. Summation technology was something licensed which Y! Can easily do themselves. Perhaps the $30M gets you lots of PR and a false sense of how cool you are (or used to be).

Posted by dyp | Report as abusive

Very astute. Right time to buy YHOO?

Posted by Geo_T | Report as abusive

John, love the wild speculation…time will tell, indexing, summarizing, snippeting, its all in a days work for search. I think your are right, the product may eventually get indexed and distill all the summarized with some Hadoop magic to trend over time.

Posted by EvanBradley | Report as abusive

What’s “Netscape”? ;)

Posted by JGillespie | Report as abusive

I am not sure if intelligent computers or search engines will actually help users. Although it seems that intelligent search engine that understands what I am looking for might be a good idea; but that can be annoying many times. Google’s auto complete feature of search phrases annoys me many times.

Posted by aambupe1 | Report as abusive

How can Summly be compared to WATSON when the only thing it can do is choose 2-4 most informative sentences from an English text? The algorithm can be implemented by any smart student with knowledge of NLP in less than a week an it shares no similarity with what’s behind WATSON.
Summly cannot ANSWER or INTERPRETATE questions, it just trims an article. It’s nothing, really.

Posted by IgorYashkov | Report as abusive

I worked for Summly.com, psycho-kid’s startup. They fired me and screwed me out of a third of my pay, since I missed a day of work while traveling to Sweden to care for my mother, who had just had a serious spinal injury.
There is an explicit clause in my contract banning me from stating that there is no innovation behind the company’s IP. The kid is a sociopath, that I am allowed to say. (They recently tried to force me to sign an agreement preventing me also from this. They offered me £50 to do so.) Brat Wonder would email everyone in the middle of the night to complain about minor bugs. In his world, all bugs are of the highest priority: ELE (extinction-level event).
The CTO kept talking to lampposts. Mostly about his . He wasn’t entirely incompetent, though, albeit a choleric Ukrainian. The CEO on the other hand, is a blockheaded bully, standard US issue, who is perpetually ‘confused’. The CSO actually understood what I said. Unfortunately, he’s a cowardly Indian, who dared not stand up to the troika.
Most of the tasks assigned to me could have been carried out by a drunken baboon. One or two might have required a sober chimp. I would say things like “This filter is pointless: the posterior equals the prior.” Blank stares from all save the Indian. “Just port it to all other languages,” would say the bully.
I did learn a new programming paradigm, though:
val = obj.field ;

if (obj == null) { … }

Posted by csamuelsson | Report as abusive

Mr. Abell, you are very far from the concept of Semantic Web. Core of idea known as Semantic Web is providing data in format that machines can process, with rules and connections to meanings. Website today may contain word “history” but that doesn’t mean it’s about history. Semantic Web would be a state where I could “ask” the page about it’s content. Technology licenced to Summly by SRI is not that.

Posted by m.domanski | Report as abusive