Okay, it’s been pretty silent around here recently and a few folks have asked “what’s up with that?” so I wanted to touch base and let you know what’s been going on with me.

WARNING: this will not be very business or marketing related, however I believe it’ll be interesting and hopefully a little entertaining so encourage you to keep reading if you have a few moments to spare.

As some of you may be aware I’m kind of an Uber-Geek who likes to tinker with everything and loves to write code. Half the tools on our PC’s and running on our servers were created in-house. In the early years it was a way to save money, if I couldn’t afford or didn’t feel fair with the licensing for something we wanted or needed to use I could always re-create it by building on top of Open Source technologies.

If you don’t need the benefit of Tech Support (as we didn’t) then why pay for Windows and MS Office on 10 workstations when Linux and Open Office will achieve the same things, and typically better? By better I mean mechanically faster and more stable without all the useless MS garbage left over from earlier versions weighing the systems down.

That isn’t to imply that I don’t like MicroSoft, in fact I love MicroSoft and truly admire what Gates and crew have accomplished in and out of business over the years. I’m just saying it was more practical and affordable in those early days to go another route. Plus it allowed me to tinker more.

Anyway, for those who know this stuff already I apologize for the recapping, I just wanted to frame what’s coming next for those who might read this and not know much about me or my business.

Over the summer I’ve been involved in some great projects, few were really profitable but that’s okay. My approach is usually that what I learn from flops is equally valuable to what I earn from the hits. I sleep much sounder since adopting that attitude.

And from one of those summer projects an idea was born for a smart application that could read (parse) Press Releases and News Stories on our topics of interest and then produce quality content on the subjects.

I’m not talking about one of those cheesy article spinners that usually just put out gibberish (I mentioned these in my last posting), but rather an application that would constantly be reading (parsing) all sorts of materials on a specific topic and in essence be learning the Who, What, Where, When, How and Why points that mattered–and which could then remember (save) that information to draw upon in creating short briefs on the subject later.

If this sounds a little crazy I understand that, but the technology already exists and is in use every day. MS Word has a neat little function that lets you summarize a document, and even my own NIMS script (or any mash-up widget) takes chunks of content on a single topic from multiple sources then repurposes them together on a central display; that’s basically what this smart application I’m talking about is doing, only instead of summarizing a single document and repurposing existing content it locates dozens/hundreds/thousands of related documents online to source from and creates an original subject brief or report.

To do all that it needs to “think” much like the way search engines do, so that instead of simply being able to focus on a single aspect of a topic, it can also associate and focus on related aspects to the main topic.

Think of it like this, if your topic was “houses” then a straight-forward summarizing approach would be for the software to just go out and learn facts about houses like:
- houses are buildings
- houses have rooms
- houses have roofs
- a house may have a porch
- and so on…

I wanted it to also find that a house “can have a garage”, and eventually that a garage is “where you park a car”… then that a car is “a vehicle people drive on roads”… then roads are “connecting pathways made by man”…

In other words, I wanted it to recognize semantic patterns to locate relationships and expand its memory (database) so that when it creates a subject brief it would have a broader knowledge base to draw from.

I want to be clear that this isn’t really Artificial Intelligence of any kind, the software isn’t alive and isn’t really “thinking” by any means. It just mines for, parses and stores data based on a large set of rules I’ve given it in the programming to accomplish the first half of its purpose. This is actually pretty basic in concept, just broad in the sense of scope.

What’s going to make it “smart” in my opinion, is when we finish the coding for how it handles the output of that stored data to create the topic briefs or reports. Again, not because the software will actually be thinking, but because the rules governing it are so intricate and complex to allow for the goal of drawing upon not just the specific details of the primary topic, but to also use pertinent details of related topics in order to produce a well rounded and articulate piece of content.

So, that’s what I’ve been engrossed with lately, and being Geeky when I jump into a project like this I tend to go all-in with it meaning other things (like blogging) get pushed aside. I experience time loss–not like those people who claim aliens abduct them, but as in it’s 3am before I even realize it every night and I have to force myself to stop for the day–but boy is it fun!

And how do I see this software being applied in my business?

I mentioned mash-ups earlier, and readers of my blog should know I love mash-ups in general. They’re fantastic for covering a topic from a central location so that interested readers don’t have to jump through hoops to find what’s important about the topic to them.

Well, as I see it, this software can take mash-ups to a new level. Traditionally, a mash-up gives you the highlights of a subject as they’re presented across multiple sources. While informative, that also has the negative aspect of being kind of stiff and shallow for some readers.

Basically, I think there are 3 types of information consumption. There are those who appreciate and prefer to read the original works of Shakespeare (traditional web sites and blogs), those who just want the “CliffsNotes on Shakespeare” (a typical content mash-up) and then there are those who want “Shakespeare for Dummies”… but where’s the web equivalent for that?

I think a smart application that can take the highlights and ultimately break them down to create informative briefs or reports is almost there. The next step will be to expand it from creating a brief to creating a brief-based discussion that stems from the topic details. And the roots for that are already in place as well with archives of discussion arenas like Twitter, FriendFeed and all those social networks out there, not to mention blogs with comments. The possibilities are really endless.

What I ultimately see is being able to say, “okay, today I want to start a web site about Urban/City Gardening” and having a software “partner” who lets me handle the technical and management details while it handles the topic research and begins the content creation.

Email this post

Technorati Tags: artificial intelligence, data mining, smart software