Diigo Links

Wednesday, May 26, 2004

85% of all statistics are lies



As a balance for my previous post:

look at the last comment on this page

Jaron Lanier points out that not only does Ray Kurtzweil handpick his datapoints, but he also manufactures his own scale. Scale is SOOOOO important a lot of the time. In the early days of surveying in Europe the Lords found the more accurate their measurements, the longer their borders. Think about it, if you measure a coastline with a 1000ft long ruler you'll miss any feature under 1000 feet. So, if you go back and measure it with a 1ft ruler you'll find the coastline is actually quite a bit longer. Lanier points out that Kurtzweil has selected the size of his rulter to fit the trends he's trying to express.

On the other hand, Lanier doesn't throw the baby out with the bathwater. He gives props to Kurtzweil.

I cooled on the idea of singularity over night. It just sounds too kooky, too sciFi. But then I started reading a CSO Magazine article on RFID (radio frequency ID tags) and I'm right back in lockstep with Kurtzweil.

Here's the comment I sent to CSO Magazine after reading the article:

One of the defining characteristics of a complex system is the inability to determine (with any precision) an initial condition from any current state. The other side of that coin is the inability to predict any future state based on knowledge of initial conditions (e.g. the weather, the stock market, etc.). With as many variables as we have in play here it's hard to say where RFID is going to go.

It's easy to imagine combining RFID with several other up-and-coming technologies (e.g. IPv6, high density ROM, spintronics, etc.) to come up with a whole new way of doing things.

Already I've seen an article on a club in Spain implanting RFID chips into patrons who don't want to have to carry credit cards. That's a simple application of a current technology that sounds incredibly "futurist".

It seems to me that the security problems inherent in RFID are in no way unique to RFID.

Security depends on rules and limits. The problem we face with RFID is the same problem posed to security in any semi-chaotic system. The boundaries of the system are not easily definable. The response will either need to be draconian in drawing an arbitrarily broad boundary around the system (i.e. lockdown, timebombs, passwords, etc) or it'll need to be more intelligent than the business processes that are created by whatever marriage of technology we see in the future. We're not yet able to manufacture "Smart" automation. So I predict what we'll see is a cry from the security/privacy advocates to regulate - at the same time we'll see risk-takers forge ahead and boom/crash through cycles of innovations and attacks on those innovations.

Either way; RFID=job security for security professionals.

No comments: