Alex Kelleher’s Blog

My first encounter with Mandelbrot was when my Dad coded it up on a Commodore 64 (yes, let the nostalgia begin).  The math didn’t mean much to me then, but the pictures were nice.  Or at least I thought they were, they looked something like this:

The thing about this type of maths is that it’s “self similar” – when you zoom in, you get the same shape over and over again, as in the animation below (courtesy of Wikipedia).  That happens in nature in everything from flowers to skin wrinkling.

So I was pretty intrigued when I found this site, which uses similar math to generate images – and suitably updated for today’s digital eye:

That’s it.  Mathematical eye-candy.

Every so often, a story comes a long within the field of psychology which underlines just how far we’ve still got to go. 

In a study at UCLA, subjects were asked to do one of 8 different tasks – from saying whether words rhymed, to counting the number of tones they heard.  So – very different, varied tasks. 

Investigators using fMRI (functional magnetic resonance imaging) – the very cutting edge of brain scanning – and then had to guess from the results of the scans which tasks the subjects were doing.

Chance alone would get 1/8 or 12.5% of guesses correct.  They managed 80%.  Pretty good – but let’s just remember that they were looking at tasks that were very (very) distinct.  And they still got it wrong 20% of the time…

How did they do it?  With machine learning math (the same kind of math that we use at Cognitive Match).   And yes, it’s a bit of a good result for this field of technology.  But far from perfect, and very far from mind reading.

Of course, all this means is that fMRI and the way we interpret the results from that technology could be better.  Much better. 

But more importantly – maybe it’s just the wrong thing to do (which is my take on it).  It’s a bit like working out whether the stock market is up, based on satellite weather images.  The brain just doesn’t work based on large areas of electrical activity on a scan.  It’s complex, enmeshed, multi-layered, abstracted and part of a body of integrated biological mechanisms.   And I hope that’s clear.

Something unusual happened today.  Something that hasn’t happened since around the year 2000.

People have started talking about a new search engine. 

Granted, it isn’t with the same “HAVE YOU SEEN HOW GREAT THIS IS?” enthusiasm that started Google’s viral growth.  It’s more of a “I kind of like the front page images”, or “you know, the results aren’t too bad” groundswell.  Microsoft, compete.com and various other are claiming good growth, and the graph I like to show every so often (which includes cuil.com and wolframalpha.com) is a good look for them (As predicted, both of the others are now on a very large downturn)

Then, something even more interesting happened, for me: I chose to do a search on Bing.com instead of Google.  And I got better results.  Yes, better results.

Okay, they were only better for the first page, and “better” is a subjective measure here, but that’s how Google caught on – by getting the first 5 results better, with less spam and less irrelevance.  Sure, Bing is subject to less “gaming” than Google is, and that might just be the explanation.  But I get the feeling that something, just a little something interesting is happening here.  It’s no revolution yet (which is a significant shame), but there’s an outside chance it’s the beginnings of one.

Great photography on the home page combined with interesting facts is fun, sure.  And good search results are crucial, that’s a given.  But MS needs to up the ante, they’ve got an opportunity now where the incumbent is feeling a bit tired and staid despite their incredible efforts to innovate.  And Microsoft could just, at a long shot, grab some market share back.  Here’s some freebie tips, Microsoft:

1. Target the home page images so they’re relevant to me 
2. Get rid of the in-context popups next to search results
3. Break out of the Google-style search results listings – a link and two lines of text isn’t very innovative
4. Given that most people (79%) click on results 1-3, make these results richer, with more information
5. What’s with the white space (on the right as below)?

Finally, to follow up on my predictions for Wolframalpha and Cuil, as they are the only other substantial pretenders from the “new” school of search.  It is a shame that they didn’t make it, but it does underline the difficulty of the task:

Building on earlier research, this piece of research suggests something that could be useful in everyday life: people are more likely to comply with requests if you ask them in their right ear.   Clearly the “club/disco” setting they used in the study is an easier one in which to target someone’s right ear than your average business meeting.  But next time you want something, maybe you’ll sit to the left of your boss/supplier/client…

Thanks to http://www.spring.org.uk/2009/06/persuasion-the-right-ear-advantage.php for the heads-up (ears-up).

In the field of realtime, one of the most exciting areas is augmented reality on the mobile.  Part of that is of course recognising what the phone is looking at.  That can be done by the G1/ iPhone 3GS in the sense that they can combine GPS location with compass and accelerometers to know pretty much where the phone is.  But by adding object recognition into the camera feed, they can complete the picture.  A pretty impressive demo below:

A great day.  VERY happy to say we’ve raised Series A Investment from the fantastic team at Dawn Capital.  I look forward to working with them and the rest of our team on building a market leader.  TechCrunch says it better than me:

http://uk.techcrunch.com/2009/07/01/hello-realtime-cognitive-match-raises-a-series-a-from-dawn-capital/