If I had the inclination to do such a thing, I wouldn't post an Ask HN about it.<p>I would:<p>1. Look at the top 50 HN contributors, looking at their profile (and possibly checking up on some of their posts) to figure out if they are already committed to some project or company that would interfere with working with you.<p>2. Auto-select any names you recognize as people you "like". You know they're smart, and not merely prolific at posting to HN. Also look into any names you know you like who <i>aren't</i> among the top 50.<p>3. Figure out your own strengths, and, based on your lists comments, narrow down to a handful of folks with contrasting strengths and interests. A team of hardcore developer that churns out code like a machine, and a generalist that doesn't mind closing vim/emacs when other tasks need doing who also happens to have either design skills or at least an eye for great design, would be my ideal combination (but this may be reflective of my own limited experience rather than any blinding insight).<p>4. Contact the short list directly, one or two at a time, with a "Hey, would you be interested in trying an experiment with me, based on RFS #X?"<p>5. Profit!
How would that work? Who would make decisions about building the groups?
Would 'good' people just be placed together - that doesn't make a good team.<p>It seems like something that could happen, but making it a community (as in HN) effort seems like it has too many issues.