This post began as a comment on Fred Wilson's blog. 

I find it ironic that Fred finds fault with Facebook's strategy which does to the web what Twitter is doing to the web. We could have a long argument about the subtleties, but they'll all be hair-splits. Fact is we're repeating a loop the tech industry has been caught up in for a few generations. We're about to enter a phase when one vendor tries to break out from the rest, the way DEC did in the age of mini-computers, the way IBM did with the PC, the way Microsoft did in graphic operating systems, and the way Google did in what we may come to refer to as the web. We're now headed into new territory, and Facebook is making a bold bid, maybe the boldest one ever, to dominate. But it won't work, it can't, they're up against forces that are inexorable. Every time through the loop the bidder gets a little further, only to fall to earth once again. 

My gut says the winning strategy will be this -- a company gets within striking distance of taking it all chooses, instead of betting the company on the unlikely event that they'll win, bets on the jungle. Sees itself more as an investment banker than a hotbed of innovation. 

I urged Microsoft to do this, before they bet the farm on Vista. I said the investment they were making of billions of dollars per year on programmers was a bad bet. Of course they didn't hear it, kept on going, and now they're right back where they were before Vista. They never figured out where you want to go today (bad joke). 

Some of the stuff in Facebook's proposal is good and worth supporting. Starting with the page-level metadata, that's totally harmless, and if it's widely implemented represents an upgrade to the web that everyone can take advantage of, including their competitors. We need to see more stuff like this. 

The next level isn't quite so harmless -- their implementation of the graph -- because all the data resides on their servers, and as with Apple and their iPhone/iPad apps, they could shut anyone off at will. This is not acceptable. (BTW, quietly Twitter has acquired the same power.) Even so, the API is admirably simple, and worth using -- but it should also be available in XML. There's lots of XML code out there, tap into it. Facebook not saving any money by not offering it and they're showing they're picky about who develops for them -- not good for trust. And regardless, their offer is unacceptable. It must be federated, the data must live in lots of places. Show us they get that by integrating the other way, have Facebook feed off graph.scripting.com, graph.us.gov or graph.twitter.com. We could all relax a bit then, knowing they're still thinking web, not vacuum cleaner. 

Any other approach is doomed. Facebook is hot now, but history has shown that being a hotbed doesn't scale. That eventually these companies have to tap into the general talent pool and they end up achieving the same level of mediocrity as the previous dominant one. It happened to IBM, the minicomputer companies, IBM again, Microsoft, now it's Google's turn, and soon it will be Facebook's. 

Zuckerberg is riding a rocket, and no doubt he's brilliant, but he won't make it. What he'll do is wake up his competitors, and one or more of them will get a little bit of their shit together, and will break out of complacency. Who knows it might be Twitter. If you see Google give up their religion and start to say no to their programming gods, and start doing things to win markets instead of glorify their genius, that'll be all it will take to put up a big brick wall that Zuck won't be able to get around. Microsoft and Oracle are still out there and either of them could create a pool of hundreds of millions of dollars to finance an open alternative to Twitter or Facebook. Even IBM is large enough to make a difference. 

At least Twitter has some ears out there. Fred and Bijan go to their board meetings. Hopefully at the next one they'll consider hitting reset on the go-it-alone strategy they outlined at their conference. A lot of the advice I gave Microsoft 15 years ago would play well for Twitter today, as well as Google, or for that matter, Microsoft.