Capturing all “facts,” redux
One of the few pleasures of growing older is watching ideas come ‘round again and again and enjoying a quiet chuckle while thinking “here we go again.” Kevin Kelly on Google+ shared a NY Times article about a company called Factual tells me the wheel just keeps on turning. Factual aspires to gather and bank all the “facts” in the universe.
Back in the ‘70 AI programmers built “expert systems” that tried to compile complete databases of knowledge by interviewing “experts” such as doctors. They ran into problems because so much expertise turned out to be “tacit” knowledge gained by docs from years of experience. It wasn’t quantifiable and amenable to database capture. The latest incarnation seems to be IBM’s Watson system that will be “trained” as the Sloan-Kettering cancer center. Here’s hoping they’ve learned a lot from earlier failures to meet expectations.
In the ‘90s came Knowledge Management (KM), the idea was that companies could capture all the “knowledge” their employees had, put it in a database, and then share it over the new corporate networks being bought. Initially KM saw knowledge acquisition as a four tier idea: 1) “raw data”; 2) “information” (aggregated data that meant something); 3) “knowledge” (information adequate for good decision making and actionable); and, finally, “wisdom” (making the “right” decision). KM is still around but hasn’t set the world on fire as promised. They’ve found, evidently, that the first two levels are pretty useless unless real people can generate the second two things: knowledge and wisdom. The Factual company seems to be starting—again—with data. Well, clean data might help avoid the old saying: “garbage in, garbage out.”
I’m not at all opposed to building on what went before to make something improved. But pardon me if I don’t get too excited (or become an investor, haha) until I see something that hasn’t been done before.