“I can also give lectures. I write the lecture beforehand then save it to disk. I can then use a part of the ACAT software called Lecture Manager to send it to the speech synthesiser a paragraph at a time. It works quite well and I can try out the lecture and polish it before I give it.”

Interesting!

>"ACAT includes a word prediction algorithm provided by SwiftKey, trained on my books and lectures"

Wow, I wonder whether they provided it before (or) after MS acquisition.

My understanding is that over the years his input setup has changed as his physical capabilities deteriorated.

This would have been circa-2004, and I want to say at the time he had a less predictive but more physical system.