Yet Another "What's the Next Big Thing" Post

23 May 2008

Reading this economist analysis of the latest OECD report on broadband penetration got me wondering something. Not about broadband penetration per se, but about productivity and its relation to computer technology. This bit specifically:

In other words, new applications that effectively harness broadband must still be developed. So far, only online media and entertainment have done this. Strikingly, this is not simply a case of people and businesses integrating broadband into their current ways, as with the PC in Dr Solow’s time. Industries like health care and education need to change as fundamentally as, say, the music and film businesses have.

The question that I asked myself is: ‘What else is there that you can only do with a broadband connection that is not media based?‘. Media is the obvious case and the big winner so far. Everything that has hit big on the internet so far because of the increase of bandwidth is a subcase of media. Music, movies, MMOs1, Skype/VoIP, YouTube, etc. All of these high-bandwidth activities are media stuff, bulk file transfers and high bit rate streams. So what else can we do with our ever faster (in some places, at least) network connections?

Brainstorming on this has proved fairly hard. Distributed Computation was the first thing I thought of. The faster and lower-latency your network connection, the amount of smarts you need to put into chunking and distributing parts of problems decreases, as coordination becomes easier and delivery of problems and solutions becomes faster. While this is pretty cool, it doesn’t speak directly to the problem of productivity in work or life. Services like Amazon’s Mechanical Turk benefit partially from the same things, but since human processing speeds have a sharp upper limit, they can’t benefit indefinitely from the increasing speed of network connections.

Thinking about this sort of thing leads pretty quickly back to the basic nature of the work that people perform and the lesiure activities they engage in, and how computers and networks can make that easier, faster and more efficient.

There are lots of people already working on moving matter around more effectively and cheaply. It’s been the number one preoccupation of humanity since communities were more than a couple of days to walk across, and computers have already the bulk of their impact there, so most improvements are going to be incremental, and aren’t going to see much benefit from network speed, as matter is a lot slower than the bits that describe it.

There are also a good number of people involved in the selling of that matter once it’s arrived (at least at an intermediate distribution center, for online retailers). There too, computers and networks have already had most of their impact. Other than more and better product information and bigger and higher quality samples, there is not much benefit from faster networks here.

For people who design things, engineers, artists and designers, there’s some network scaling to be had from collaboration to be had in terms of collaboration. Unfortunately, this space is hard to break into, because it’s much easier for Autodesk to add networked collaboration to AutoCAD than it is for someone else to design and build a collaborative drafting and design application from the ground up.

For certain types of collaborative office work, there are also some scaling benefits to be had. Google is already doing some interesting work in this space, and I’m sure that there are at least a couple of fortunes to be made in it (although at some level it feels crazy to go up against Google at this point, especially when they already have such a lead). Nor am I convinced that there is a great deal more network scaling benefit to be had here. There is a low upper limit on the useful complexity of office software, and it’s already fairly painless to deliver these applications to the web browser.

What strikes me, at this point in my thought process about this issue is how little I know about how other people spend their time at work. To have any useful ideas about this, I think that I’m going to have to have a better grasp on how many people do what and what their jobs actually entail. A better idea of how productivity is measured and what it actually means would help, too.

I figure that I have to be missing something. Otherwise there’s not a whole lot of reasons to keep increasing the amount of bandwidth that homes get much beyond the point of being able to stream 2-3 HD programs and along with a couple of phone calls and some bulk transfers. That’s a big connection, but not that far off. Of course, it’s quite possible that these new, major uses for networking are things that no one is doing yet, and that no one sees the value of at this point, in which case, I’m unlikely to be the one who thinks of them first. Still, trying is an interesting exercise.

1) MMOs have experienced another benefit: being able to smoothly handle more other players in your immediate area at once. I consider this a subcase of distributed computation.