There has been a lot of crap going around recently among emergent people and their sidekicks - unscrupulous, sometimes plainly ignorant journalists - about the idea of GoogleNet. This magical thing is supposed to replace, or at least supercede, the Internet. The base for these assumption is the small-scale testing of a Google-run WiFi network, but more importantly, the fact that Google has recently been releasing apps to facilitate every aspect of normal Web usage.
The assumption is that because Google has in the past become the default-choice search tool, users will stay infinitely loyal to the brand and continue to stick with anything Google-branded.
In reality, Google only does one thing well. Google is often touted as the anti-Microsoft, a people's champion to defend poor users against The Beast. In reality, the two companies are quite similar. Just as Microsoft gets its money from selling Windows and Office, and uses it to finance the Xbox and MSN, Google gets its money from advertising and uses it to pay for its forays into niche markets. The reason it has become the default choice search engine was because it worked so well. In the time of "portals", Google presented an empty page with a search box, and a bunch of text results out the other end. Add to this the fact that it really did find the most relevant results, and you will understand the key to the whole phenomenon:
It just worked.
Whereas the rest of the nascent GoogleNet doesn't. Gmail may give you a lot of storage, but you don't need a lot of storage; they can say thay they'll give each user 2.5 gigabytes of hard drive space (and more) because the overwhelming majority of users will never come close to filling up all that - especially since Gmail includes half decent spam filters. The whole minimizing letters and conversation management business is very neat, but it only works in Internet Explorer (people's champion, huh?) and anyways, for most users this does not make up for the fact that you need to wait for ten seconds once you log in for the whole mess to load.
Google Maps? Fancy, and maybe useful if you're in NYC, but I'm in Small Country. Google Earth? Not even my office workstation has the muscle to handle it properly, and that's got a gig of RAM. Google Talk? I haven't bothered yet, but the sources I tend to trust say that it is objectively inferior to both MSN (killer feature: you can see when the other person is typing something) and ICQ (killer feature: you can send messages to people who are offline, and they'll get them). Actually I haven't bothered with Google Talk for the same reason I haven't bothered with Skype - nobody I know uses it. ICQ got popular because it was the first one to market and MSN got popular because everyone with a Hotmail account tried it; the killer features don't help you get users, only keep them.
In fact the single useful thing Google has done in recent memory was to add the calculator to their core search function. So now I keep going back to google.com every time I get in a flame war on a car messageboard, because it allows me to easily convert Euros into Canadian dollars, pound-feet into Newton-meters and miles per gallon into liters per 100 kilometers (that last one is particularly tricky, because the values are inverse, so there isn't a simple coefficient to multiply by).
To build a GoogleNet, Google needs a strategy that would get customers to use a Google app for everything - browsing, email, IM, etc. Unfortunately Google is an online company, so they are building online tools. This is not the best way to do things.