Cameron’s right – we need to be working toward @font-face (that’s Kernest’s mission), EULAs need to support @font-face usage. But Cameron’s also wrong on 2 points – First, @font-face support really only exists in Webkit-based browsers right now (the big win comes with support in Firefox 3.1). Second, @font-face doesn’t require .eot files, ttf works as well 
(Embedding faces for use in Internet Explorer does require .eot files – this capability has existed for more than a decade.)
The biggest problem I see with Cufon – is that it is proprietary. Font-embedding within browsers through proprietary means has existed for 10 years now – with little to no adoption. Like the web itself, for adoption to take off, it needs to be an open standard.