NEWS: Semantic Web investments

Ed: Semantic web needs to solve data tags, inference, and user query interpretation. AI tricks to capture email, web pages; distinguish companies, people, place mentions; and tie with data stores like Wikipedia. Little discussion of the inference that provides value-add.

Semantic Web investments

One of our primary investment themes over the last couple of years has been around the semantic Web and "Data Intelligence".  ..

I have frequently mentioned Zoominfo as a portfolio company driving substantial innovation in this space.  Zoominfo can understand unstructured text to compile aggregate summaries of related information occurring across the Web.  By understanding that certain words on a page mean a "person" or a "business" (i.e., smart data), Zoominfo can impose structure around this information, enabling search and information discovery in ways that far transcend the capabilities for plain keyword-based search (such as Google, for example).  ... (Ed: service for marketing, recruiting to displace LinkedIn.)

Initial Experience with Twine

Today Nova Spivack has taken the covers off of Twine and he will demonstrate it today at the Web 2.0 conference.  My prior post on knowledge networks was an attempt to lay the predicate for the concept of Twine as a knowledge network.  So I'll now make this more concrete to explain how I have been using Twine to build my own knowledge network during our initial alpha period.

Knowledge networking is what we VCs do in real life.  I am constantly asking

• "Whom do we know that knows about this" 
• "Didn't we talk to someone a year ago who knew something about this?" 
• "Isn't company X a likely customer of company Y and what do we know about companies X and Y"
• "Didn't [partner z] do a reference check that referred to this new technology some time ago?"

Like everyone, we make heavy use of the web, often sending to each other the  things we have found that might be of long-term interest or relevant to a current project. Within the partnership we communicate largely through email and then we meet every Monday to sync our activities and processes.  Organizational memory is largely in our heads.

My use of Twine is trivially easy.  I use two 'on-ramps'.  One is that I bcc: my Twine account with all my emails.  (I also set up some rules in my inbox to forward copies of certain emails to Twine as well, but my email rules are not as smart and extensive as Twine.)  The other on-ramp is I use the Twine bookmarklet as I browse with Firefox.   These two methods capture pretty everything I  consume in my business life.

As the emails come into Twine, Twine enriches the email to find companies, people, and places in the body. These enriched links connect to other content I have captured as well as Wikipedia and other sources.   So now I have a thread I can follow of what else I know, read, or can find that is relevant to answer the questions above.   This is less critical in the moment than it is  days or weeks after I  receive the new information.

I do the same thing as I browse.  When I find something of interest, I can immediately "bookmark" it into Twine and associate it with the things that I track in Twine.  (I have a personal account and am a member of several Twine  groups, including the Radar Networks Board of Directors group.  Other members of the group see the same information I have place into the group's Twine account. 

This is the real power of Twine.  All the information I have been accumulating has been intelligent interconnected as a personal semantic web of knowledge.  That mimics my own ability to recall what I know.  I don't find this that interesting, yet, because my use of Twine is relatively recent.  Ten years from now, my long term recall of what I knew today will be greatly diminished, unless I use Twine.  More immediately, everyone in my groups (Crosslink Capital, Radar Networks, etc.) benefits from attaching their knowledge networks to mine.  This really allows us to create a group diligence process that represents and leverages everything we, as a firm, know.  It means I know can truly leverage the knowledge and relationships my partners accumulate. The enrichment means we all get more than we put in as we use the product.  This basic process of structure knowledge capture and sharing is nearly universal in business, from major account sales processes to product design collaborations.  We all know that email is fundamentally broken as knowledge capture, retention, and sharing tool.

There is a lot in Twine I don't use, to be honest.  Photos, videos, and threaded discussions are all part of the application.  I don't need this, today, though perhaps in the future.  And there is a lot I'd still like to see in the application, including support for  meeting creation and calendaring synchronization.

The challenge with Twine is discovering all the consumer and business use cases and bubbling them to the top.  But for a terminal early adopter, I have to say it's really going to become an important enhancement to the way I communicate and accumulate what I know (and who I know.)

This is going to be a slower rollout than most web applications.  There is a lot we don't yet know about how to best package it as well as people's usage patterns.  The computing machinery behind the curtain is substantial.  So we still have a lot to do to understand some mundane but essential things like what it costs to support a user.  So please be patient.  The private beta is likely to last quite a long time until we get this right.  We are rapidly learning to live by Thoreau's guidance to simplify, simplify, simplify.

10 Semantic Apps to Watch

Written by Richard MacManus / November 29, 2007 

One of the highlights of October's Web 2.0 Summit in San Francisco was the emergence of 'Semantic Apps' as a force. Note that we're not necessarily talking about the Semantic Web, which is the Tim Berners-Lee W3C led initiative that touts technologies like RDF, OWL and other standards for metadata. Semantic Apps may use those technologies, but not necessarily. This was a point made by the founder of one of the Semantic Apps listed below, Danny Hillis of Freebase (who is as much a tech legend as Berners-Lee).

---

In September Alex Iskold wrote a great primer on this topic, called Top-Down: A New Approach to the Semantic Web. In that post, Alex Iskold explained that there are two main approaches to Semantic Apps:

1) Bottom Up - involves embedding semantical annotations (meta-data) right into the data. 
2) Top down - relies on analyzing existing information; the ultimate top-down solution would be a fully blown natural language processor, which is able to understand text like people do.