[eruleman]

"My hope is someday, when the next Aristotle is alive, we can capture the underlying worldview of that Aristotle - in a computer. And someday, some student will be able not only to read the words Aristotle wrote, but ask Aristotle a question - and get an answer!"- Steve Jobs, 1985

[wwweston]

"What bothers me is, nothing does." - Dixie Flatline in William Gibson's _Neuromancer_, 1984

Wonder if Jobs had read that by 1985.

[dr_dshiv]

Really? We just made a version of GPT3 finetuned on the complete works of Plato. It produced some solid new dialogues. About, for instance, the relationship between beauty and the good.

[andrepd]

> We just made a version of GPT3 finetuned on the complete works of Plato. It produced some solid new dialogues.

I highly highly doubt that.

[ekianjo]

the mind of a person is ever changing. you cant capture it across all their works.

[ehecatl42]

And how was GPT3's Attic Greek? Did you run it by classicists?

[trevoristall]

This concept is explored in-depth in the Hyperion book series. Highly recommended if you like the genre :)

[jonathanstrange]

Maybe he should have tried reading some of Aristotle's books.

[somenameforme]

A slight tangent, but there's something extremely interesting on the topic of Ancient philosophers related to this Jobs' quote. Socrates wrote nothing down, to the point that some have claimed he was illiterate. And that is a possibility, though improbable.

But the reason he claimed to not want to write anything down is because, in a nutshell, 'books cannot defend themselves' : words can be taken out of context, meanings misconstrued, and text made to mean what the interpreter wants to make it mean instead of what the author meant for it to mean.

It's quite fortunate for the world that many of his students disagreed, but it's interesting nonetheless.

[imranq]

This is probably the best use case for this technology. Although I think there's a lot of potential for very realistic fake news

[ekianjo]

You dont need fake news to be realistic to be believable. You just need to repeat them over and over again. Works wonders.

[nomagicbullet]

This is such a great idea. And for me it goes even beyond being able to exploit the intellect of deceased geniuses. It could be the path to generating believable AI personas out of everyday people. If sufficiently accurate know I would love to have an AI with my father's advice and sense of humor once he is no longer with us.