So I'll summarize quickly.

1. Swat, those are great ideas! I could definitely stream on Twitch.
2. I think people will learn more if they do more of that core work, but it comes at the cost of things slowing down. If I get us started and up to a point where people can pick their own favorite tasks to work on without stomping on other people's toes, people might have more fun and we'll get more done. So those are the tradeoffs.
3. Maybe we should have an optional "architecture planning meeting" the evening before or the morning of for an hour or two that would allow people to be more involved in those early stages.
4. Maybe it's broader than just architecture. Maybe we can all jump into a Google Doc and come up with a basic functional design doc, prioritize the features we think are most important, and people can start thinking about what tasks they'd like to work on, in addition to coming up with a plan for core architecture and the design of the core classes.
5. Maybe we just extend it to 10 hours with the first two being that architecture stuff, and have it be optional.
6. Maybe we just keep it at 8 hours, and just assume the first 2 will be planning/architecture, and just live with it.

Just thinking out loud here. There are no real commitments to anything yet. If you've got feedback on any of this, or any alternative options to throw out, bring them up!

I don't know what the solution to that particular problem is yet. Maybe I should step back a bit and not try to build the core. Heck, maybe I should step way back and just plan on helping people with their problems, as well as being available for design discussions, etc. and not so much writing a lot of code myself. Maybe it gets solved by having more smaller teams. I don't know. I'm open to suggestions…

I was actually really happy that you did the core work, as it prevented me from needing to do any of it and allowed everyone im sure to pick a "task" and know immediately how its supposed to integrate with the game.

That being said I'm not sure that this problem is very "work aroundable".. As usual I Don't want to pressure you into doing it but it is probably the biggest job of the game and you are the best/most experienced person here. (or atleast.. we all assume you are :D)
I have a few suggestions..

1. Since this is a learning experince for most people anyway.. You could twitch yourself doing the source code and explaining how it worked and what you were thinking as it went along, so as everyone watching (and on voice chat) could learn something and as soon as your done you wouldnt need to do very much explaining before everyone is on there merry ways.. Then.. after that's done you could go help different people with there problems or continue with the core.

2. If we are still assuming that we will need someone to do the core and that the core should be done before others get there tasks, you could start programming 1 hour in advance.. or the day before to get the minimum of the core needed before others can start working.. and since you would have an early start.. you could prehaps finish early and watch/tutor others thorough the second half/4th quarter of the challenge.

These are my suggestions.. and I know nether of them really take a lot of pressure off you.. but at least others wont be twiddling there thumbs and infact might learn more from the result.

But I wanted to sort of outline what we did last time around for people curious about how this collab thing might play out.

Keep in mind that everything is changeable. I'm quite certain we'll want to change at least something this time around. So we'll just have to figure it out as we go. But at least this is a starting point.

So here's basically what happened…

1. A few weeks in advance, we had people sign up. This way we knew who would be joining us, and people could clear up their schedule.
2. We picked out a game to clone a few days in advance. (I still highly recommend we pick a simple game or part of a game to clone. It will make life easier for us.)
3. A few days before, I did a crash course training via Twitch on how Mercurial/TortoiseHg/BitBucket worked so that we could all share code more easily.
4. The collab was officially scheduled for an 8 hour block on a Saturday, and since there were four of us and we were all in the USA in relatively similar time zones, we were all able to just work together without any real problems. If we have people from Europe or Asia or something, we'll need to have multiple teams or something. I don't know right now, but we'll figure it out.
5. We all got onto the chat room here on the website, as well as a voice chat server. Voice chat was nice, but that didn't work for everybody so we still did a lot in the chat room here.
6. I spent the first 1.5 hours of our 8 hours trying to put together some basic core code while everyone twiddled their thumbs. I'd like to fix this somehow.
7. I spent the next 1 hour explaining how that basic core code was intended to work, we came up with a list of tasks to work on, and people all picked something that they wanted to work on.
8. The next 3.5 hours went much better, with everybody working on their own little piece, then moving on to the next piece.
9. Our 8 hour time block ran out and we stopped for 20 minutes and did a little retrospective on what worked and what didn't work.
10. Some people had other stuff to do, and so they left at this point.
11. Some people kept making changes all evening as well as Sunday and into the week.

I actually think I had more fun making this game in this collab than I did in the month-long competitions. It was fun to work with other people on these things, and the very short deadline meant (a) we had to focus to get stuff done, and (b) that we were free to just be done with it when the time ran out, without a massive commitment.

I'd definitely like to find a way to make sure that everybody is actively doing stuff the whole time. The first couple of hours wasn't a lot of fun for people because everybody else was just kind of sitting around doing nothing, while I felt a ton of pressure to hurry and get to a point where others could do something useful.

I don't know what the solution to that particular problem is yet. Maybe I should step back a bit and not try to build the core. Heck, maybe I should step way back and just plan on helping people with their problems, as well as being available for design discussions, etc. and not so much writing a lot of code myself. Maybe it gets solved by having more smaller teams. I don't know. I'm open to suggestions…

If we've got people from around the world, we'll have to come up with a way to make the different time zones work. That might mean running two projects 12 hours apart, and letting people pick between the two (or doing both). We'll have to figure that out too. (But that's a good problem to have.)