Before I start, I should note that I haven't worked for Linden Lab for more than a year now, so I'm not giving away any inside information, nor do I have any inside information. Yes, my opinions are formed partially by what I learned while I was inside Linden, but I'm writing this based on my observation as an interested outsider.
I think often the way to kill a business is to over-monetize it. I remember the 1990's, and Web search engines. The pattern was repeated over and over again. There'd be one that was the best. They'd realize they were the best, and they'd either get sold or they'd try to monetize their business. The page would go from being relatively clean, to being a cluttered mess of ads... and the search results, being increasingly paid, would become less and less useful. So we'd all move on to another engine. That ended with Google, who had the vision not to try too soon to over-monetize their search, and who recognized when they did monetize it that they had to do it in a way that didn't completely undermine what brought people there in the first place. If you're doing something new and cool, you're better off focusing on how to make it cooler than you are focusing on how to squeeze a buck out of it.
How does this apply to Linden? Many think that Linden's best asset is their technology. It's not. Don't get me wrong, the technology is good; but it can be replicated. Server side, OpenSim is already well on its way to replicating (and in some ways exceeding) the server side technology of Second Life. Client side, the Second Life client is still way ahead of anything out there that's not based on the GPL Second Life client source, but that won't last forever; after all, there are lots of MMO clients out there, so it can be done! No, Linden's greatest asset is its audience.
I know Philip got this. Talking to the company while I was still there (a year before I left? I don't remember) about OpenSim and competition in general, he said that OpenSim and interoperability would only be positive for Second Life. After all, Second Life had the greatest number of residents. Even if that didn't stay, a growing interest in and a growing usefulness of virtual worlds would only help Second Life. Alas, the company as a whole didn't realize this. What they should have been focusing on was promoting virtual worlds. Instead, they... well, to be honest, I'm not really sure what they were focusing on, but they didn't direct substantial effort towards promoting virtual worlds in general. And that was the mistake that, I believe, led to the massive June 9 layoffs, or what I call "Lindenmageddon".
In 2008, a team of engineers from Second Life working together with a team from OpenSim premiered a bold and exciting experiment. From the beta grid (Aditi), users were able to take their Second Life accounts and teleport to regions running the OpenSim server. The functionality was extremely basic, but it was a concept demonstration. This was the first step to interoperability. I was only peripherally involved in this myself. As a member of the server release team (perhaps server release manager by then?), I helped the engineers working on this get some regions set up, and would warn Whump whenever we were going to do things to aditi that might interupt his operations. (I also annoyed Whump a few times when doing things on the beta grid without giving him enough warning....) But, when the experiments were happening, I joined in with other Lindens and residents to see what all the fuss was. I felt like I was walking on the moon. I remember commenting to others that we were exploring grand new worlds, taking the first (and very limited) step in what was going to be a gigantic new thing. People outside the lab were talking about the possibility of true interoperability in the main grid (the production Second Life environment) during 2009. It was all very exciting. Second Life was bursting out of its shell, and was finally providing the seed for the true global metaverse that it had always promised that it could turn into.
Alas, after that, the interoperability work coming out of Linden Lab slowed to a trickle. It didn't stop, but outside nobody else saw anything new. Many of the engineers working on it were redirected to work on the (now defunct) "Second Life Enterprise" product. (I always personally thought that was a dead-end, but I was careful who I said that too. I was annoyed that some release people we needed for other things had time taken away working on it, and I was sad that the manager whom I loved, Josh Linden, got pulled over to it. But I was also sad to see the company working on the 3D version of Novell Networks instead of the 3D version of the Internet.) Yes, Linden still did things; Enus, at least, was working on the PyOGP stack, and Zero and Infinity were involved in standards meeting and developing the WVRAP protocol. (Incidentally, Infinity left Linden Lab a month ago, and Zero was part of the June 9 Lindenmageddon layoffs.) But there was no big effort. Out of this lack of more public progress from Linden was born Hypergrid, Crista Lopes' working interoperability protocol between separate OpenSim grids. Hypergrid still has some things it's lacking, but it got much farther than any implementation out of Linden.
And now concurrency (number of people online at one time) in Second Life has been sagging, and the lack of growth has caused financial distress for the lab. (I don't have inside figures, but Tateru Nino figured out at least this much from looking at the public figures.) Yes, it's a recession. But Linden made the mistake of having people in control who were business types, who think in terms of consumers and monetization. They didn't get Second Life. (Truthfully, nobody does, and the first step to understanding virtual worlds is recognizing that you don't get it. It's a tremendously flexible potential platform. Like the Web, if you try to put it in a conceptual silo and productize it, you're going to be making something like early 1990's walled-garden AOL, which will fail in the face of the true Web.)
I work with MICA, the Meta-Institute of Computational Astronomy. We maintain a presence in Second Life, even though really we'd rather be working on open source servers that we control ourselves-- i.e. OpenSim. There are two main reasons why we stay with Second Life. The first is that OpenSim doesn't yet have a good Voice solution... but that will change. The second is that Second Life is where everybody is. The popular talks that I and some others give wouldn't make sense anywhere else, because the audience isn't there. And this is what Second Life's greatest asset is.
If Linden Lab had focused on helping make virtual worlds take off-- make them more useful by providing functionality people wanted and needed, working on interoperability so that people could take their Second Life accounts to and from software that was developed not only by Linden engineers, but by everybody-- I predict they would have done a whole lot better. Their already existing audience would have given them a leg up, and would have kept them a leader or at least a major player. Yes, they would have been helping "competitors", but by raising the profile, utility, and popularity of virtual worlds in general, they would have helped themselves.
I predict that the decline of Linden will set back the adoption of virtual worlds several years. And that may not be so bad. After all, before the Web, there was Gopher, which was never all that big. Before that, there was CompuServe, a walled garden that a lot of people used, but which became increasingly irrelevant as there was the Web. Second Life had the opportunity to be the seed for the 3D web, but instead they chose to focus on being the 3D CompuServe. As we go into the future, OpenSim, or another platform that defines its protocols, provides a working reference implementation, and supports true interoperability will form the basis of the true global metaverse. And, once we have that, and once people are able to develop cool things to run on top of it as flexibly as possible (i.e. not limited to just LSL and renting space on Linden's servers), we'll see the metaverse take off. As to whether or not Second Life is part of that, we can only wait and watch. It will take more vision than trying to somehow connect to current hot buzzwords ("social media") if they really want to be a part of this. And even if they are, they may not be a meaningful part. (After all, AOL is still a part of the Internet, is it not? But with the possible exception of the AIM protocol, they aren't a meaningful part.)