I think open threads are in practice already this.
Not that I have noticed. Open Threads seem to primarily be "here's a cool thing I'd like to let you know about". If I want to post something like "The 'you are cloned and play prisoner's dilemma against yourself' example against CDT is actually pretty bad. Solving it doesn't require UDT/TDT so much as self-modification, with which even CDT would be able to easily solve it." (with a few more lines of explanation), for example, my model of Open thread predicts that if I'm wrong, I'll be downvoted a few times, and may or may not get good feedback. Also that Open Thread is meant for things that are more of interest to everyone, rather than being fairly specific. Which is why I'm not posting anything like that, even though I'm 80% sure that particular example is correct and may be of interest to at least some people.
Excessively encouraging such things could breed crackpots.
In what way? I doubt existing visitors to Less Wrong would be significantly more likely to generate crackpot ideas because of the existence of a thread, and I doubt even more that more crackpots would come to Less Wrong to participate in one thread. It may, admittedly, reduce conformity if people find unexpected support for non-mainstream ideas, however I'm not sure that most would consider that a bad thing.
my model of Open thread predicts that if I'm wrong, I'll be downvoted a few times, and may or may not get good feedback.
I think downvotes would depend on how you present your idea. If you present your idea as if you're already convinced you're right, and you're not, I think that would lead to downvotes. But if you preface your idea with "hey, here's something I thought of, dunno if it works, would appreciate feedback," I think that would be fine. What people respond negatively to, I think, is not wrongness so much as arrogant wrongness. (Or at...
If it's worth saying, but not worth its own post, even in Discussion, it goes here.