Apex Muse is intended to liberate developers from traditional development constraints, in search of better ways to turn raw data into useful information, something from which you can build a story.
Muse as we see it fits an interesting niche. On the one hand you have the huge nanny runtimes like the JVM and CLR, which are trying to become more dynamic, even supporting special "runtimes" within their virtual machines to enable scripting languages.
On the other hand, you have the native-compiling languages that also are trying to become friendly to scripting, without becoming bloated and overbearing from a developer freedom point of view.
Muse is an attempt to strike a balance between these two worlds. The reason it is separate from Athena is there are two high-level categories of languages: statically typed imperative languages and dynamically typed functional languages. Trying to cram both of them into one runtime is the mistake of both camps to date.
Muse will always focus on, and specialize in, dynamic functional programming. Languages and language features that seek to exploit the benefits of that paradigm will find Muse to be a natural runtime, grown organically by bringing the best of three very popular languages with dynamic and functional programming support.
Over time we see adding more languages to Muse's repertoire, and implementing Muse on multiple platforms. We also see Muse incorporating the best ideas from some of the most intriguing runtimes that have been proven in the distributed computing space over the last 30 years.