Very interesting note by John Robb about the role of the 14.4K modem in the rise of the Internet and Web as we know them today. He makes some excellent points, provoking thoughts on two points. The first is to look at why the Web succeded in a 14.4 world, while other things fail with similar capabilities. One great example in the US is the total bust over the last few years of web-like data services to wireless phones. Based on the 14.4 modem story that John recounted so well, you'd think these things were poised for success. Most of the cellular data services had about that much bandwidth to play with, but with such limited UIs it was really a wealth of bits compared to a desktop system. And sure, most of the phones required a separate call to support the data services, but the modem had to do that as well (we all have that sound pattern from the negoiating modems etched in our brains). It sure felt like the recipe for getting a new wave of services off the ground! But it wasn't that simple. Right before this was taking off I had an executive of one of the wireless companies tell me "We are going to make sure we aren't going to be turned into dumb pipes like we were with the Web". Translation: the end-to-end architecture of the Internet kept us from setting up the kind of toll boths and controlled on and off ramps that we wanted; we're going to make sure this new highway systems comes with those things built in from the beginning. The result: they squashed any kind of innovation by controlling the services via controlling the service and application providers. The services that were rolled out were the result of small bunches of people strategizing in comfy conference rooms, not the Darwinian result of trial and error of huge numbers of people innovating at the edges. The rallying cry became "we need 3G with it's extra bandwidth and always-on capability", meanwhile missing the point that they already had better networking capability than when the Internet started it's meteoric rise. Why am I highly confident in this theory? Because the one service that did take off in the world, iMode at NTTDoCoMo did allow that kind of hacking – and it worked for them. Now everyone here is trying to emulate it – we'll see if they're willing to open up the highway enough to let it happen. The second point I come to from John's discussion is to accept the rise of services that are a blend of servers in the cloud and massive computing on my desktop, but to then worry about the increasing trend of the bandwidth providers to limit the ways in which I use my bandwidth. As Lessig and others have clearly articulated, the end-to-end model that was undone in the cellular data services is under fierce attack from home bandwidth providers (DSL, Cable modem, etc). The only reason that I can see that this isn't more of an issue, is that we've been successful disguising these new applications as more web traffic, which these folks have decided they have to live with for the moment. But I ask you this: if something as wildly successful as the web starts to take shape, are these folks going to sit back and allow themselves to be "dumb pipes" again? Said another way, if the web started today, would they let it happen, or would they setup tollbooths or try to block the traffic in the name of some social benefit? I believe in the kind of apps that John talks about, I only worry that if they get too popular people are going to take notice this time.