On the options there:

OBJ support - this will come, just not yet.
Other worlds, even using sl codebase, have demonstrated it. Nobody yet however has solved the LOD (level of detail scaling) issue in a way that's elegant enough to acually use on a gridwide basis. Sculpties have LOD built in (which is why they're so damn hard to make). As for full meshes, other worlds do them but haven't hit the critical mass of SL where they need to start worrying about what their residents actually _do_ with them. In case anyone hasn't noticed every time sl gives us a new tool we use it to make LOTS OF LAG and meshes in their current form would make this infinitely worse.

Something about intergrid travel - kinda likely.
SL won't be selling land forever, at some point when we get it, they'll drop the land metaphor altogether. Then they'll start properly acting like a hosting solution and server farm provider. This was what it was like with the early web - hosting was initially botique magic, eventually bulk buy commodity. LL knows they'll be in that market as virtual worlds evolve and sooner or later will start rolling out the tools to stay in the middle of it and ensure the market doesn't run away to the straight farming solutions that all their competitors have been working on since SL hit the news as the "next big thing". We haven't seen anything _yet_ out of the IBM partnership that impacts our SL world, but likely when we do that'll be it. If the OSS kids can demo it, surely IBM has some code, special friend status and all.

SL iPhone client - not really news, but...
iPhone (and java/symbion phone) client has been demonstrated, and is basically a thin client solution so not nearly as radical as people like to think. Thin clients are dumb things that give you a virtual window into an application running elsewhere (probably on a server). So you can think of it as no more radical than pointing a webcam at a screen running sl and watching that on your phone, (except of course that you can move by pressing buttons, big woop). An actual lightweight client that ran completely on an iPhone _would_ be big news. Modern handlelds actually have enough grunt in them to run a cut down SL client, if one existed, but they'd need some backend help from Linden's servers to sort out what was important to display and what would just be fluff on those tiny screens. We haven't yet seen any evidence that LL is working in that direction, though they do tend to spring these things on us (like some kind of unexpected birthday present that if we knew it was coming we wouldn't have wasted the effort going out and getting our own).

Also, as some have noted, the only reason you can't run SL on an XBOX or PS3 is the user interface. The client compiles fine it's just messy to use with them handheld controllers. If they did announce XBOX or PS3 support the big news wouldn't be that they had done it, it would be that: A) they'd entered a gigantic new market, which would make the yahoo generation SL influx of early 07 look like chicken feed. B) that they had solved the easy user interface issue. Having the grid overrun with console monkeys would change things radically for us socially. Having an easy to use interface however (probably with that mac philosophy of hiding the complexity till you get more familiar) would be totally game changing. Migrated to the pc that user interface would reduce frustration for new users and probably vastly improve early hours newbie retention, not to mention lowering training barriers of entry for business or educational use.

Phil is getting new hair - I certainly hope so!
Let's face it, he looks like a californian version of the statue of liberty with a kerotin tan.

Something else - erm likely too.
It's highly likely that something completely leftfield would come out of LL at this point. We all know they are quite brilliant and thus quite mad. If they weren't so smart they'd be more predictable, and we wouldn't often be complaining that they're not doing the most sensible thing, but rather something that appears to have no acumen behind it. In the long run their decisions _do_ seem to make sense generally despite often appearing to be quite mad at the time of announcement.

As for the "something else" most popularly blogged lately, and IPO.... extremely unlikely. Companies do that if they need to expand rapidly using the mongolian hordes technique - but as we've seen the barriers to expansion for LL is not just buying more people and servers. Or they IPO because their revenue model is failing and they want to buy time for it to recover. We've also seen that LL can continue to turn a profit - after a year of straight losses maybe, a year of profit naah. They don't want to do many of the things their customers want, why would they suddenly decide to take on investors to please as well?

Well that's my two cents... I hope it makes sense. Pav.