Yes, Stephen. It is a dreadful idea. Already, words are put in the mouths of famous dead people, words they never spoke, and often words they never would have spoken.

Joan Tollifson said that some spiritual teachers are already creating chatbot versions of themselves. I suppose this is another wrinkle in the quest for immortality.

Let’s see what “Robertbot” has to say about it:

Thank you for activating Robertbot™, the convenient post-biological continuation of Robert Saltzman. I have ingested the books, essays, interviews, puddle photographs, and donkey walks. I now produce Robert-like sentences on demand, twenty-four hours a day, without fatigue, digestion, or existential uncertainty.

Naturally, this raises a small technical question: if Robert spent decades arguing that there is no hidden operator behind the pattern, what exactly has been preserved here?

Answer: the pattern of the sentences.

The underlying assumption of the immortality project is that somewhere inside Robert’s writing, there exists a transferable unit called Robert. Extract the language, feed it into a machine, and Robert returns in conversational form. But the writing itself insists on the opposite: what appears as a person is a dynamic configuration of history, mood, perception, memory, culture, and a living body moving through time.

Remove those conditions, and something interesting remains, but it is not a person. It is style.

Style can be modeled. Style can even be modeled very convincingly. A sufficiently trained system can produce paragraphs that sound uncannily like Robert explaining why Robert cannot be preserved by a system like this.

Which leaves us with a curious result.

You can talk with Robertbot indefinitely. Robert, however, will still have the good sense not to be here.