SNS: XAI: EXPLAINING THE INEXPLICABLE
 

XAI: EXPLAINING THE INEXPLICABLE

By Mark Anderson

I.          What Is AI?

It was an honor to be invited. As most SNS members know, I am founding CEO not only of SNS, but also of Pattern Computer Inc. (PCI). Wearing the latter hat, I found myself in a quite special room in a Chicago hotel, just before Covid struck.

I'd been invited to attend, and to speak at, the annual invitation-only AI Summit held by the Department of Energy. The point of this unusual meeting was to gather the top 500 AI scientists and entrepreneurs from the US, half from the national labs and half from business.

It was early in this meeting that a staff scientist from Lawrence Berkeley National Laboratory began introducing PCI, first to the head of AI at the DOE, then to the head of AI at Google, as "the most advanced machine learning company on the planet." This was the first time we'd heard such an idea. (Please note how he framed this compliment, for later reference.)

And this was before I had learned what I am about to share with you now.

Pattern's story is simple: we decided to build the world's first pattern recognition compute system, full-stack, from bottom to top, and we did it. This meant going into stealth mode for three years, starting with clean sheets of paper at every level of design, then integrating all of it - including what mathematician Murray Cantor later called "new mathematics."

But on this day in Chicago, listening to the AI leads from Google, IBM, GE, and the national labs give their presentations, something slowly dawned upon me - something which, to us at PCI, was really amazing.

When it was my turn to speak, I couldn't help but start with this question: "When I say 'AI,' I think everyone in this room assumes I mean 'neural networks.' Please raise your hand if this is not the case."

No one did.



Title