Saturday, February 24, 2007



-- Jack Krupansky

Knowledge mail and knowledge messages

One interesting idea that I would eventually like to pursue is a concept I call knowledge mail. The basic idea is that rather than manually composing text messages for our online communications, we need tools to compose knowledge and to compose queries about knowledge. The twin goals would be to simplify communication and to enable intelligent software agents to participate at a high level in our online communications.

I do not have any of the details worked out. I just know that so much of what I do via email in an ad hoc manner could be codified and structured and automated. I do belief that it is a very hard problem, but one worth pursuing.

If anybody has reason to believe that my faith in this concept of knowledge mail is misplaced, lets hear it.

Open question: How might this blog post itself, which I submitted via an email message, be represented as machine-readable "knowledge" rather than ad hoc prose?

I suppose I should use the term knowledge blog or knowledge blog post to adapt the concept from mail to blogs.

Or, maybe we should use the term knowledge message to generalize the concept to encompass email, blog posts, instant messages, etc. I think that makes sense.

Another open question: What problems would need to be solved to enable and provide support for online communications based on knowledge mail and knowledge messages?

-- Jack Krupansky

Sunday, February 11, 2007

Intel Tests Chip Design With 80-Core Processor

I was thrilled to see the news that Intel was testing an 80-core processor. Not that such a chip by itself will be terribly useful for much more than research, experimentation, and niche applications, but it heralds a bigger leap into what I call ultra-massively parallel computing with thousands or even tens of thousands of simpler processors on a single chip, and a form of computing hardware that will finally be able to enable hard-core use of software agent technology where you have many thousands of lightweight processes running within a single system.

The computing infrastructure of the future will be based not upon very complex cores, but very large numbers of simple cores.

Nonetheless, this news of an 80-core chip is a truly exciting advance.

-- Jack Krupansky