I was recently a guest on Falken’s Maze, a cool sounding name for an even cooler podcast about semantics and AI.
In case you only know me through my writings at Mashable and aren’t a long-time friend of mine, you may not know that I’m a huge buff of AI, and have had a lot of experience programming and configuring various chatterbox systems (my particular favorite side of the AI field). I spent a lot of time in the mid 90s as well as after the tech bust in 2000 playing with, coding for, and engineering various chatterbox systems.
As I said on the show, as I’m getting acclimated to where AI is currently after having put it down for a few years, I’m surprised by several things. First of all, I was surprised that semantics and semantic web is actually, in part, about AI. It has a huge marketing problem. When you hear it described, you typically get the impression that it’s all about tagging.
The semantic web folks, unless they want to stay stealth, need to push the AI side of things. Everyone loves Minority Report and I, Robot. Associate cool stuff like that with your brand, and you too will be cool. Simple as that.
The other thing that I’m impressed with is that “the cloud” hasn’t yet been suitably applied to chatterbox AI. The major problem with chatterboxing is the fact that you need immense computing power and an immeasurable database of responses to really beat the Turing test effectively and consistently. We have that now.
I haven’t seen evidence of this train of thought applied yet (and maybe that means I need to “get on that project”), but I’d love to be proven wrong.