Clay Shirky "Industry Perspective" Impressionistic transcript by Cory Doctorow doctorow@craphound.com July 8, 2003 Supernova Conference Washington DC -- Minitel: Hahahahaha! It's near Jul 4, we need to make fun of the French. The French govt underwrote the development of a network through France. When you've lived in no connectivity and ANY connectivity shows up, it's a great leap forward. However, once the system was up and running, managed centrally, innovation flattened. The PC started to catch on, but MiniTel was connected. Then modems, but Minitel could guarantee QoS. Then the Web: the fusion of the value of the Internet and the value of the PC, and Minitel was no longer the right answer. And Minitel had cost them an enormous cycle of innovation. They overinvested in soemthing that was a good idea for a long time, and once it became a bad idea, they couldn't see it -- the overcommitment cost them the period of innovation. We've got it worse than the French do: as big a screwup as Minitel was, ours is worse. We're making our mistake in the physical layer, in twisted-pair. It's the bizmodels that say, "It oughta be a good idea to run a circuit-switched, voice-optimized network" -- that was a good idea for a network. We missed the packet opportunity: we missed the opportunity for flat-rate. Bob Lucky turned down the opportunity to manage AT&T's Internet twice. If the situation is this dire, why are more people not up in arms about it? We're habitual techno-optimists. When engineers look at the network, they perceive a center and an edge. Connect a device to the "edge" and it can talk to the rest -- bellheads and netheads both see this, though they disagree about how the connection should work. If you believe this, life looks good: you can walk into Best Buy with a grand and walk out wiht Gigaherz, Gigabit Enet and Gigs and Gigs of HDD. Soon we will have 10Ghz Enet and TBs of storage: phenomenal. When you look at the center, it's mindboggling: we've run enough fiber to go to Jupiter and back, but we haven't put any of it in my basement. There's tons of fiber, AND we know how to put more and more data in it. In a edge/center world, this kicks ass: take your cheap box and plug it into your cheap backhaul and you've got a 1000X improvement -- it's literally unimaginable. It's not just a fast PC and a fast network, any more than email is fast IM. But we've got THREE entities, not two: 1. Center 2. Edge 3. Phone company Phone company: people engaged in monopoly rent-extraction for the copper in your basement. We distinguish long-distance from local. Creamy fiber center, crispy twisted-pair crust. Anyone running a NAT at home is running a small and fast network in their homes. The Edge isn't one device anymore. There's highspeed in the home, highspeed in the center, and a brittle line in between. It's a dumb-bell-shaped Internet, fat in the middle, fat on the edge, skinny in the middle. I got Vonage, and three months later I cancelled my second Verizon line, cancelled Sprint, my LD carrier, and then Verizon was forced to lower the cost of its DSL. For every dollar of IP they sell me, they lose $2 of inflated voice service revenue. It's a crisis for them, and it'll get worse once you can buy 100Mbit pipe. You can't simultaneously protect the incumbents and create new ways of getting fat pipe in the home (fiber or wireless). There's no bizcase for that. That's the American Minitel: if you protect the incumbents, you kill innovation (which, in this room, is like mulching your lawn with ground-up baby kittens). The likelihood of any solution being deployed is in inverse proportion to its quality. The telcos have realized that "If we do nothing, nothing will happen." -- There are only two ways out: 1. Functional competiton: Lots of ways to buy unadulterated bandwidth in the home displaces incumbents (powerline, WiFi, municipal fiber). I'm not optimistic about this. 2. Change the regulatory environment. Hard. Phone companies are as close to regulators as the family farm is. Regulators can always find a spare $180BB for the telcos. The telcos bought physical plant -- switches -- on the idea that it was going to last for 10yrs. This is like buying a PC with the intent of having it last for a decade. Dumb. Option 3: 3. Stagnation. We're not #2 behind Korea on broadband -- we're NUMBER 11. Not even in the top ten. We're falling (Isenberg: we're #15 now). Even though Sean Fanning and Claude Shannon are both Americans, we're #15. (Isenberg: we're below 15 -- the scale only tracks top 15) We don't have the solution to the dumbbell Internet. We can't muster the political will to connect the edge to the center. We could do what we did with cellphones: pick the wrong tech, pretend we didn't for a decade, and then switch at great cost. And the rest of the world will have lapped us.