I've been round this loop now three times, at least. If you count the big ones. There were small ones too. 

The big ones: 

1. IBM. 

2. Microsoft. 

3. Google. 

When IBM hit the wall, it was with a revolution they called the Micro-Channel Architecture. It was touted as a way to take back the PC industry from the cloners. But it was also a way to reign in the power of Microsoft, who was IBM's upstart. Didn't work, it only cemented Microsoft's position, though it took Microsoft a few years to realize it. 

With Microsoft it was the great call to arms in late 1994, when Bill Gates rallied his team and told each of them to maneuver their battleships and aircraft carriers into position. He thought he had met his own Microsoft (he had been waiting for it) and its name was Netscape. Not realizing that the problem wasn't Netscape, it was a sea-change in the tech business analogous to the one that IBM had failed to overcome. His upstart was the web, not Netscape.  

Now it's Google's turn.  

When IBM hit the wall, I had an idea for them. It was, interestingly, the one that launched my career as a blogger, back before we knew it was called blogging. I wanted IBM to pick up Mac OS, which was then languishing. These were the years that Steve Jobs was out of Apple, starting NeXT. I felt Mac was the right product, but Apple lacked the management and resources, the gravitas to compete with Microsoft. I thought Mac was IBM's ticket to get out from under Bill's thumb. 

When it was Microsoft's turn, I had a better idea of what would happen to them, and urged them to kick back and find ways to profit from the growth of the web. I felt they could get the largest share of the growth by just being the background, making investments and providing services. They didn't agree but I didn't think they had a choice. And that turned out to be correct. 

Similarly, I don't think Google has a choice. Their "social" offerings have been rebuffed repeatedly, and they will continue to be rejected by users, no matter how promising they are, no matter what they are, different from Facebook, a Facebook clone, doesn't matter. Why?  

1. You can't make revolution with employees. 

Can't be done. They don't know how to do it. And if they did, there's another problem: 

2. Everyone's watching. 

So on Day One your service pretty much has to be feature-complete, and ready for hundreds of millions of users. Forget about corner-turns. Forget about dipping your big toe in to get a sense of the temperature. These are the advantages of the upstart, when they're starting. People have responded to yesterday's piece by saying basically that Facebook can't rip up the pavement any more than Google can. True. But their innovation is done. Now they're reaping the rewards. But when they face their upstart, they'll be in the same place Google is in now. 

The crazy thing about #1 is that you'd think that Page, like Gates before him, would know that his employees don't know how to make products that overturn the status quo. Why? Because these guys do know (or did once) what it takes, because they did it. What they need is to find people, who, like themselves, are good at iterating and bad at taking no for an answer and determined to make their place in the world. It's too late for Larry, he's already made his place. It has to be someone new.  

But if you were to, by luck, find such a person -- if you hired them, then #1 kicks in. Because the guy who can turn it around isn't good at fighting internal BigCo political battles. He or she has no patience for it, and it's not what they do well. It's very easy for the Gundotra's or Horowitz's (who are the best at BigCo politics) to push them aside as trouble-makers or incompetent, one who doesn't play well with others.  

Aside from that, there's no test you can give someone that determines the probability of him or her being the next Gates or Page. We don't understand the qualities of such a person well enough. And since luck matters too, you really need to launch a thousand ships and let the best one win. Or maybe none. You can't be sure how this experiment will turn out. 

Google should take an inventory of its skills and figure out which ones could help the new Bill, Larry or Zuck be successful. And be the one to provide it. My guess is that Google has a huge lead re Facebook in infrastructure and ability to deploy. I bet it'll be hard for Facebook to overcome that lead (just as it was impractical for Microsoft to get into the PC-making business). So if I were Larry, I'd make the cloud to end all clouds and price it really cheap for any entrepreneur who's willing to stake their future on being the next Big One. And when one or more emerge from the pack, buy them lunch and ask if you can invest. Bring flowers, and candy too.  

Funny, it is, that there is someone doing exactly this.