Aaron’s still at it, fixing bugs in bundler, nokogiri and maybe still teaching even. I’ve stopped using MRI because the toolchain and ecosystem aren’t conducive to economical scaling or profiling like C or JVM. JRuby is much easier to profile using JMX, but it’s harder to get native extensions going and not the same. For backend, lately I’ve been more in the Elixir/Erlang, Rust and Go camps for solutions.
Depends on what you mean by the roadmap, but TruffleRuby looks very promising. I haven't used Ruby in years, but if I did return to it TruffleRuby would be the first thing I'd explore using.
When Apple had some interest in Ruby and things like RubyMotion started to appear, I thought maybe the Dylan lessons would be of inspiration, but it was in vain.
