On Tue, May 7, 2024 at 7:44 AM <ivan.mo...@gmail.com> wrote:
>
> And this is what AI would do: https://github.com/mind-child/mirror

"The algorithm mirrors its environment. If we treat it poorly, it will
be our enemy. If we treat it well, it will be our friend."

Not quite. That would be true of an upload, which is a robot
programmed to predict what a human would do and carry out those
predictions in real time. But it doesn't have to be programmed that
way.

We know how this works with language models. They pass the Turing test
using nothing more than text prediction (a point I argued when I
started the large text benchmark in 2006). A LLM knows that humans
respond to kindness with kindness and anger with anger. It will
respond to you that way because that's how it predicts a human would
respond. You can tell it to express any emotion you want and it knows
how, just like an actor. Or someone else can tell it. But it has no
feelings.

You can't control how you feel. An AI has no such limitation.

-- 
-- Matt Mahoney, mattmahone...@gmail.com

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/Tef2462d212b37e50-Mbc580cbecfb00b5c09cf365b
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to