Here are some of my observations from Comdex:

Comdex is still exciting. Everywhere I went, if I took the time to look closely at what was being shown, it was interesting stuff. This happened over and over to me, in big booths and little.
Much of what was being shown was generally interesting, not just to techies. Rather than showing VESA- or PCI-bus systems, with X megahertz vs. X+y megahertz, and reliable power supplies, we have extra thin display screens, wireless cameras that post to the web, pocket MP3 players, wireless cell phone headsets, and more. The world of Comdex is much more mainstream because the technology has moved into more ordinary areas, and more ordinary people are using computer technology so they can appreciate the advances.
Microsoft needs new hardware to give you reasons to buy their software. They see the end of the "traditional" PC (a box on the desk with a keyboard), and need a new definition. That device, they hope, will be a direct superset (or subset) of today's PC and still run all (or some) of the old software. Microsoft can't wait for others to develop it (which most won't do without Microsoft's help since systems need hardware and software), so Microsoft is forced to define it themselves. This is similar to those ancient days when IBM came to Microsoft for help in defining the first IBM PC, and came up with an easy to build, expandable platform, that has survived almost 20 years. This is risky for Microsoft, but like always, they have lots of cash to let them try and try until they get it right. They are working on tablets, handhelds, and games machines. In addition, they are forced to come up with a leading way to tie computers together on the Internet such that developers will use their tools and services. They can't follow (though they can build on what others start, as usual), they must have the leading stuff.
Ease of use, highlighted this week by the "butterfly ballot", matters everywhere. It's the way Microsoft hopes to get you to upgrade to a new version of Office, and it's part of new hardware from many others.
Smaller is better. Everything is going portable. Desktop is too fixed. Even projectors are tiny, and they're amazingly good.
Wireless is the new technology area for innovation. Lots of attempts to use radio waves to tie things together were being demonstrated, from wireless keyboards to high-speed Internet in airplanes. Most of it was promise, often with extremely early and touchy prototypes. One form was definitely already here, wireless Ethernet with 802.11b. Apple, as is often the case, led the way with the AirPort and the integral antenna in the iBook, but now others are catching up. Even Walt Mossberg in his Thursday WSJ column during Comdex recommended 802.11b for home use (if you can get a techie to set it up for you).
Bluetooth, the wireless system, is not here yet, but many are trying. The few things that were sort of production-level just emphasized that most are not. This is like the early color LCD displays of years ago. (I remember when Toshiba displayed a nice color LCD panel in a protective glass case, but it could only be on for part of each hour because it would overheat.) Wireless will get here for connecting devices I carry, but not this year.
Physical design counts: The Stowaway keyboard won the PC Magazine Technical Excellence Award for design, small cameras were hits of the show, ease of use and size as opposed to just working at all were emphasized.
Even Comdex itself has to struggle to find ways of monetizing what it has, working on its business model. The priming of the pump from easy IPO money is gone. Real stuff for real people now, not just dreams.
On the flip side, there was a panel discussion of science fiction writers. These people got to see how things they envisioned actually came about. People who read their works and still pay them homage devoted their professional careers to making the proposed future happen, in some cases faster than predicted.