Paul left a thoughtful comment about my post about data, my life, and finding pieces of same. His point is that our ever-more powerful computers still suck at tasks a 4-year-old could easily do (like pick a photo of mommy out of a collection of other random photos). Google works well on the Net, where it can mine the human intelligence encoded into content links, but does less well on my computer HD where no hyperlink hierarchy exists. Most corporate intranet searches are miserable for the same reason, IMHO.
Many very smart people are no doubt spending a lot of time thinking about this,: Google alone, no doubt, has the PhD equivalent of the Manhattan Project thinking about making search not suck, and not just for text, but the places it really sucks like photos and video.
I’m wondering if we’re looking at this all wrong. Computers are decades from mimicking any sort of human intelligence (if you believe Ray Kurzweil): what is it that they can do well that might be more useful to us? The simple smart keywords in Yep, coupled with Acrobat’s OCR already have made p2 (aka the Paperless Project – I really need to write up a better description page) far more useful than drawers or shoeboxes for holding incoming statements, notices, medical documents et al. Indeed it’s incentive to keep pushing paper through the ScanSnap. But that’s the point: it’s simple things that machines can do that bring a benefit.
People I think highly of have encouraged me to look at DevonThink, a database-cum-AI engine, which I’m currently in process of doing. I’m also looking at Outliners and other tools, including human management systems (courseware), but I still think a new approach is begging to be found…
{ 0 comments… add one now }