Tuesday, August 26, 2008

Vinge on the Future

One of my mantras in my HCI classes is that often the best interface is no interface - essentially automating it away by doing the task or making the decision in the software which may or may not require AI.  A New York Times article today explores Vernor Vinge's speculations on where this automation may end.  The article refers to Vinge's essay on, "The Coming Technological Singularity" and his new book RAINBOWS END.  Vinge speculates/extrapolates that by 2030 computers will become so juiced that a new intelligence will emerge that will surpass humans and then what do these new intelligences do about us - thus begins the Post Human world.  That "then" has been the fodder of many SciFi books and movies (e.g., the Matrix Trilogy).  It also has been discussed in books by folks such as Kurzweil's, THE SINGULARITY IS NEAR, which expands on Vinge's essay.  Exploring this train of thought does cause one to give a sideways glance at the block of silicon on your desk!

My own take on this, hardly novel, is that it has been so difficult to get software to do anything that it is a lot farther off than 2030, if ever. In fact most of what I teach is focused on trying to coax software to doing anything intelligent and at least approximate what we want. Of course the humans involved in the software enterprise, do not make it any easier!

Vinge's potential solution is to partner with the computer and become better together.  That is appealing since each entity has unique skills to bring to the table and this seamless interactivity between human and computer is something we strive for in HCI.  In fact HCI as partnership is something to explore, that may lead to new way s of viewing the user experience.  Hmmmm more on that later - and more on emergence.

By the way, Vinge is a great sci fi (and computer textbook) author.  I would highly recommend sampling his work.  I often refer to his story, "True Names"  in first nailing the issues on cyberspace and privacy - he is one of the most prescient of the current scifi writers.

I would appreciate your thoughts on this and I thank the NY Times science section for the article.  A nice thought generator for these dog days of August!    Later!

Thursday, August 21, 2008

A new era of "help" messages

Often in my HCI classes I state that the best human computer interface is no interface, that you should automate the task completely if you can.  I just came across a pointer to a video from doggdot, a blog aggregator, that is worth posting and you can find it here.  It shows a robot correcting a human in a cooperative task.  Two points worth noting in the video: (1) the dialog design that results in a rather firm correction to the human and (2) the fact that the human had to adjust the platform so the robot could place the bolt on the wheel - there still is a long way to go before you will work on a daily basis with an obnoxious robot.  Gee the aibo is discontinued only to be replaced by obnoxious robots!

I think I have my nth wind and hope to increase the frequency of my posts going forward.  Later!