The technical parts of doing something like this are fairly easy. I went from knowing nothing to having a podcast within a couple weeks once I decided to do one. The hard part is driving it. Whoever takes the helm has to be prepared to put in some hours. I was surprised that producing a 30-min episode takes 8+ hours. I'm not sure if that would be less for an unscripted podcast, but it still won't be nothing. Warn your SO you'll be taking on a long-term project, and be prepared for them to be irate for several weeks as they adjust.
I'm not sure if that would be less for an unscripted podcast, but it still won't be nothing.
Sure, rather than 8+ hours it takes you about 1 hour... plus the decade or two that it took you to get your ad hoc pseudopublic speaking and domain knowledge up to the level where you can produce that amount of content at approximately the same level of quality without previous preparation.
Would it be possible to have a monthly podcast on Less Wrong topics? A possible format could be roughly four panelist (maybe half core and half rotating members) discussing theoretical and practical aspects or rationality, AI/singularity, cognitive science etc.
Episodes can be also easily framed by assigning some reading from sequences or recent LW articles and then discussing them in podcast form. This format seems to work great for www.partiallyexaminedlife.com (quite entertaining and informative podcast albeit on the diseased discipline of philosophy).
To keep things interesting occasional episodes could be done in the form of discussions with guests (via skype), e.g. the usual suspects from the SAIA and fellow AI scientists, people like Robin Hanson, Aubrey de Grey, other rationalist/skeptical bloggers/podcasters but also AI skeptics and so on.
The level of the podcast should be still accesible to newcomers but the discussion could thread into bit deeper waters. I would love to hear discussions also on more technical topics (like CEV, Solomonoff induction, AIXI etc.). Just imagine how exciting could be discussions on the more controversial decision-theoretic paradoxes!
Further possibility is from time to time plan ahead and cover in more depth the topic from that week's Harry Potter and the Methods of Rationality podcast, which would potentially help to increase the audience (it would be like a post-grad HPMoR).
What do you think? Could this be done? With the depth and breadth of material we have here and all the interesting people to talk to I don't think there would be a shortage of topics.
Edit: Changed the (bi-)weekly timescale to monthly.