You make a decent point here. But not reaching out and starting the conversation means you miss out on the only group in the world I know of who has successfully emulated animal connectomes in software and robotics. There is PLENTY of press showing and proving that including an early video seen by over a million people. So I gotta call BS on that point.
There is real world observable proof including OpenWorm that Tim cofounded and was critical in transferring scientific knowledge into software. So again I have to say BS.
Onto the general intelligence claim. None of us have ever said our work is done and putting our feet up. The first question that must be answered is “what is a general intelligence benchmark”? If someone has one for us, please let them step forward. ImageNet ain’t it. And so we’ve had to go out and find a child’s test that we’re working on now. We may or may not publish our results. To our customers, though, definitely.
The coffee problem. Maybe. And have talked to some folks about demonstrating that. I believe we can do far better than any mathematic system with very little training data. But that isn’t going to be a customer solving problem so why should we feel the need to validate ourselves to outsiders when we can work one on one?
Just like narrow AI has a broad definition, so does General Intelligence. We are at the beginning, not the end.
Viv claimed it. They got snatched up. We know some folks at Numenta. We also know a few investors in some of the other companies listed by the CB Insights article. Do they show off their tech including the robotics one? Nope.
I’ll tell you what I tell everyone who asks. Email me, show me your project is real, big, you have willingness to pay, we’ll get NDAs in place and start talking about a POC. Then you’ll get to see first hand the power of Biologic Intelligence over narrow approaches.
We are more general. That’s the point. And we will show our work in private, not in public.
from Stories by Sean Everett on Medium http://ift.tt/2pxTz7Y