At this point spam is not really an issue. I currently manually moderate all comments that come into the system. Obviously this isn't going to scale for too much longer but it was the quickest way to move forward. In the long term when spam becomes more of an issue (because it inevitably will) I'm planning to allow paying users to input an Akismet API key.
If we grow out of that I can implement my own bayesian database and teach it to filter spam.
The sky is the limit in cat and mouse game of spam. I'm just waiting for the mouse to become smarter then what I currently have in place. As for (re)Captcha systems, as a user myself I find them dreadfully annoying and it is a major design goal to not use them if at all possible. Obviously if pushed into that direction I might change my opinion about the matter and make it an optional setting.
The other option which is already in place for paying users is to prevent comments from showing up until explicitly approved.
Leave the first remark and get the conversation going!