The talk is really about using Erlang in a hardware multimedia switch. I briefly mentioned CORBA within the talk because I spent about 16 years working on it, and was comparing it to Erlang.

The real news is that he fell in love with Erlang. And mentions how screwed up CORBA is.

Don't get me wrong, we did a lot of good work in CORBA and learned a lot while doing it, people still use it, it's still very widely deployed, and a ton of interesting research was done based on CORBA. But one of the reasons I think so highly of Erlang is that it treats distributed system concerns as first-class citizens, whereas CORBA was born of RPC notions of trying to hide distribution behind programming languages. Now, it's not entirely as black-and-white as I just stated it, but that's one of the fundamental differences. Erlang just gets a whole bunch of things right.

I didn't get that far in.