If I heard it correctly, Phil's basic argument was that the flaw in SRTP was that the keys had to be carried inside the SIP signaling session, which is usually wrapped in TLS/SSL on a hop-by-hop basis, meaning that any server along the path could grab the SRTP key before re-encrypting the SIP messages and forwarding them off to the next hop in the path.
Phil noted that his encryption approach, called ZRTP, separates the encryption for RTP from the SIP signaling path, and also eliminates the need for a PKI to support large-scale deployment.
Phil mentioned his Zphone source code based on ZRTP would be available in early March, and wouldn't in and of itself be a softphone, but rather a shim that would work with anyone's SIP-based softphone to intercept and encrypt RTP streams. Alan Johnston, one of the authors of SIP and a co-contributor to the Zfone project noted on the VOIPSA VoIPSEC mailing list that he and Phil are currently finalizing an Internet Draft for ZRTP for submission to the IETF later this month. Alan argued that there are many problems with SRTP that are addressed by ZRTP.
I suppose the question is given the market momentum behind SRTP, do the vendors feel that it is "good enough" or do they see a real benefit in converting their products to use ZRTP?