> Now, stop imagining and realize that this is what has
> been made possible by inexpensive, high gigabytage
> hard drives, improved search algorithms and relatively
> cheap, yet remarkably powerful CPUs.
A couple of years ago I was with some relatives by marriage -- old money, family trust fund people -- and they were talking about investing in some sort of new database technology that was supposed to be 10000 times faster than what's available currently. I expressed skepticism -- sudden advances come in tiny increments, not the sort of leaps you see in movie macguffins. They assured me they had had it checked out with some very established specialists, pioneers in databases in fact.
The whole thing puzzled me: advances like that just don't happen without huge theoretical breakthroughs that aren't kept quiet. But these aren't the sort of people to be taken by grifters -- aside from being intelligent, worldly and very secure, I would expect sheer social insulation to provide a certain amount of protection.
Then I realized that I hadn't been applying my computer science background properly -- marginal algorithm efficiency can scale exponentially in both good and bad directions, and, with large enough problems, give you the sort of advances that were being discussed. I corrected myself by saying that such improvements were possible if the databases were enormous, and they confirmed that was the sort of application this was for.
For such offhand comments about investments that will be forever out of my reach, the knowledge has sat there like a hood ornament.
-- Andy