I’ve been hearing for years about technology that can break down the sound of a person’s voice into an array of vowels and consonants and digitally assemble them and make that “voice” say anything. I was hearing about this 15 or 20 years ago. Roger Ebert’s talking “Alex” is cool for what it is, but he needs to sound like himself. There are thousands of hours of tape of him talking. It can’t be that hard.