by Jason Preston on March 18, 2009
From Ars Technica, on Microsoft Live Search:
The thing that really stuck out to me is when Weitz said, “Search is not solved.” Since I use Google, I immediately thought this didn’t apply to me. After all, Google was god and it always helped me find what I was looking for. Weitz was indeed referencing Live Search data Microsoft had started drawing conclusions on, but I quickly realized that there were broader implications of what he was saying: this conclusion applied to web search in general.
As I write this, Steve Broback is trying in frustration to find and download a driver he needs. You’d be surprised at how useless searching has been.
Search is not solved.
by Jason Preston on March 17, 2009
A lot of people on Twitter are passing around the link to Jesse Newhart’s (justified) little rant about how Facebook is trying to be Twitter. I know this because I looked it up on Tweetmeme – goodbye Techmeme?
I think everyone’s noticed recently that Facebook, in it’s recent home page redesign, is starting to look a lot like Twitter (or more accurately, like FriendFeed).
The interesting thing, though, is that when I first joined Twitter in 2007 I kept calling it “Facebook Status Updates Plus,” because that’s what the functionality is. So in a way, Facebook is really just trying to catch up to its own potential.
But I think the real problem is that Facebook, and probably all the other social networks, will ultimately follow AOL into walled-garden obscurity. Facebook is an exoskeleton, and they know it.
They’re so damn worried about Twitter not because they want to be the hot new thing, but because they can’t let open, platform-level technologies siphon user activity out of their black hole. Once that happens, they lose their leverage and their value as a walled garden.
by Jason Preston on March 12, 2009
Have you read Dave Winer’s post about Twitter and their “Suggested Users” list?
I think his title is a bit melodramatic, but he’s essentially correct about the relationship that Twitter needs to maintain with its users: hands-off.
Unlike Winer, I can *completely* understand where they’re coming from with a feature like that. Twitter, after all, is kind of a weird concept to people who don’t yet get it, and following some interesting users is a great way to get into it.
I guess the problem is the hand-picked nature of the list. What I might try, if I were working on the next feature, would be this: ask new users if they want to go through the “suggested user guide.”
If they do, provide a series of questions (check boxes, maybe?) about their interests, and then generate a custom list of potential users based on both their filled in profile information and some basic word-use analysis on their tweets.
It won’t be perfect, but it would be fair. I doubt even Dave Winer would complain about that.
by Jason Preston on March 11, 2009
This afternoon, Steve is presenting at the SDMA Market to the Max conference, which means that I’m lurking around the building taking notes and…making blog posts. The conference looks to be a great event, and of course we love the SDMA, so you should click at least one of the links above 
Here are a few notes & thoughts prompted by the keynote, Debra Aho Williamson of eMarketer.
[click to continue…]
by Jason Preston on March 10, 2009
by Jason Preston on March 10, 2009
As many of you are already aware, we at Parnassus have events and conferences in our DNA, which is why we can’t help but get excited about 140 | The Twitter Conference.
We just took the site live a few hours ago, and we’re still fleshing out the conference sessions, but we already know that Alex Payne, API Lead at Twitter will be giving a keynote.
I’ll also repeat here the call that Steve made at the top of our schedule page:
We are actively looking for case studies focusing on app development or the use of Twitter in corporate or political campaign environments. Email steve AT parnassusgroup DOT com if you have intimate primary knowledge in these arenas.
If that sounds like you, please shoot him an e-mail. We’re looking to put the best possible people on stage for 140.
Of course, if you’re interested in attending, you can get your tickets here. Remember, first 35 tickets get 30% off…
by Jason Preston on February 26, 2009
The Economist article Social networks: Primates on Facebook discusses theoretical capacity of human social networks and cross references to real-world data.
Robin Dunbar, an anthropologist who now works at Oxford University, concluded that the cognitive power of the brain limits the size of the social network that an individual of any given species can develop. Extrapolating from the brain sizes and social networks of apes, Dr Dunbar suggested that the size of the human brain allows stable networks of about 148. Rounded to 150, this has become famous as “the Dunbar number”.
And then the correlating data from Facebook:
The Economist asked Cameron Marlow, the “in-house sociologist” at Facebook, to crunch some numbers. Dr Marlow found that the average number of “friends” in a Facebook network is 120, consistent with Dr Dunbar’s hypothesis.
Note the phrase “social graph” is nowhere to be seen in this article…
by Jason Preston on February 26, 2009
About 18 months ago, I had an epiphany, and connected some technological “dots” that I didn’t feel anyone else had put together, and threw some money at developing the idea. I had conceived a process for delivering accurate analysis of sentiment. It seemed to me that a series of linked services could conceivably take us way beyond the abysimal 60-70 percent accuracy we currently see with typical automated NLP (Natural Language Processing) systems.
Lo and behold, 9 months later we had a working prototype that was significantly more accurate at tagging for sentiment than what all the big-time, VC-funded systems were offering. We’re talking 90/95% + accuracy. Feed in a thousand URLs that link to blog posts or articles, provide a product/brand name, and it quickly spits out each URL tagged positive, neutral, or negative in relation to the product/brand provided.
I showed this to Jeremiah as soon as it was complete, and while I sensed he largely saw us as a “me too” product (I don’t think he quite grokked my table-ponding accuracy claims…) he was gracious enough to connect me with the proper analyst at Forrester who deals with such products for a demo. Between his schedule and ours, we have yet to present. That being said, I am more motivated than ever to arrange a time to show off what we have.
Why is that? Because Jeremiah just said in his recent post about Lithium’s Insight Report, that (emphasis mine:) “Although we’ve yet to see great tools to glean opinions, demonstrating qualitative information such as quotes and even sentiment will be the next step.”
The next step is here Jeremiah. Just download this spreadsheet of the URLs we analyzed for sentiment regarding the MacBook air. Check the accuracy, as it beats any other automated system on the planet.
Or you can review the report we created based on this data. (The quotes Jeremiah desires are a part of what the system can extract as it tags.)
