I learned a great deal from that piece of software. It practically begged for modification... and I modified as much as I could fit in memory... and then a little bit more. Looking forward to hearing this!
After hearing the interview, I think it was probably Carnival that I had most of my experience with.
I also wanted to agree with him on BBS Express and I'd throw Oasis into the same bucket as well. Those assembler-based pieces of software were pretty rock solid, fast, and reliable. But they were for the people who had a fair amount of money but weren't interested in creating any kind of customized destination. At the time, I called it "a war of the look-alikes" because, in my area, Oasis was extremely popular, and all the Oasis boards looked and acted completely alike. It really sucked the creativity/exploration/wonder of visiting BBS systems.
I do remember seeing the functionality he described with Carina, though. It had a cool I/O route which made the OS believe that whatever was coming in from the modem was a local keyboard input, and any text going out to the screen was also sent to the modem. So you could write a normal program which uses PRINT statements to write to the screen, and you could accept normal input as if the user was sitting right there.
That was totally awesome because it made programming telecommunications software super-easy. The downside, from what I saw, is that if/when a user was able to break out of your software (due to a hardware fault, bug, or whatever), they got access to the BASIC interpreter to do whatever they wanted. They could literally type anything on your machine. So it made your BBS potentially vulnerable. I'll assume that he eventually found a way to mitigate that.
I'm going to have to look over a copy of Carina to make sure.