1. Ultimately, No. Wires supply bandwidth to wireless endpoint devices. Small cells = more bandwidth. This means you're going to want a wire in your house, just like you have now. There may be a transitional phase where you are routing backbone traffic across a wireless device, but eventually you're going to want that bandwidth locally. This is especially true in urban areas. Rural areas may need less in the way of wires. You may see rural wired telecom go away.

2. I want a device that streams over wifi and is a phone and is a PDA and is an mp3 player. Streaming will provide a short term IP solution. VoIP will happen because I don't want to also carry a phone, so my wifi phone from my house is also my wifi phone outside my house is also my ipod...

3. WiFi sort of. More bandwidth will be allocated. It will be more controlled. It will look a little like cellular ultimately, but more open. You'll have the same sort of AP in your house that your cellular company has out in the street. Small cells win. Platforms you can innovate on win. Using the same network card in my house, office, and on the street wins. 3G is a high power big cell solution that doesn't work for everything. You will PAY to use your neighbor's wifi. The networks will charge a flat rate and pay people who run access points a metered amount.

4. I have no idea. In some cases yes. In other cases no. This is basically how telcos generate cash to pay down debt; let the local government eat them... It will depend on how well the telco managers deal with their debt problems, if they can be dealt with at all.

5. The cost of dealing with this on top of the current debt load could really kill the RBOCS. I have no idea about Cisco/Lucent/Nortel.

6. I don't know on the technical side. However, I think the IETF is too dogmatic for its own good. Its totally subverted by the vendors and cannot see its own flaws. It may become irrelevant very rapidly as running code tends not to be produced there anymore. The market makes you interoperate, and having a standard is as easy as publishing a document. Interoperable standards always win. If people need QOS, the networks will build it. The networks will eliminate IP spoofing and solve the relay problem. These will be features that the router companies offer the ISPs/updates to sendmail. You will need to explicitly tell the network if you want to provide a service, and your OS will update automatically every night. You may do MPLS tagging on your desktop. The networks can enforce MTUs. It will be "IP" but it might look a hell of a lot like ATM and the network will become as smart as possible in an attempt to avoid commoditization of the service.

Now, who the hell should I invest in? I have no idea who is going to make the right decisions here, and there are many people in positions to do so. Shame I'm not one of them.

Some thoughts on that last post...