The front page of Lavabit proudly claimed "so secure even we can't read your email." Either they really believed that, which means they have trouble understanding how security works, or they were being disingenuous. I'm not sure which is worse.
You're correct, Lavabit worked within the confines of existing email technology. A standard email service is vulnerable in three places: the service owner can choose to access user email, anyone who compromises the service can access user email, and anyone who compromises the SSL connection from the user to the service can access user email. Lavabit was vulnerable in exactly the same three places, despite all the cryptography handwaving.
In any other light, anyone in the security community would have laughed the Lavabit design out of the room, or written it off as just another snake oil security product. It's really not even a question, and I'd be interested if you could find a single well recognized contributor in the security community who would say otherwise.
Ladar is trying to do more handwaving with really embarrassing explanations of his "secure memory" system (insanely ridiculous and again, really troubling), but doesn't address the fundamental question of whether his primary security claim ("even we can't read your email") was true.
Let's also not forget that when Ladar complied with the government's request and provided his SSL key, the entire history of his users "secure" emails were compromised.
So here's the situation: we have someone who was either previously disingenuous or failing security first principles who is asking for funding for a followup endeavor that we have no technical details for, and are supposed to trust that he will develop appropriately. If the only metric we have to evaluate is Ladar's past technical performance, it seems like caution is in order. Because last time, 400k users were compromised unnecessarily.

http://www.reddit.com/r/IAmA/comments/1qetvk/i_am_ladar_levison_owner_and_operator_of_lavabit/