A simulation of me: fine-tuning an LLM on 240k text messages

Re:A simulation of me: fine-tuning an LLM on 240k text messages by jazzyjackson on Thursday January 04, 2024 @08:37PM (#38874663)

I also get the vibe that digital cloning will be popular. Maybe some extremists will think its unholy and some addicts will lose sense of reality, but for the vast majority of people, I think it's just a user interface - maybe to some particular piece of information in the chatbot-as-librarian role, or puzzle boxes that eventually reveal information once you ask the right question, like Will Smith in I Robot.

What I don't like is that the better it gets the more a radical contigent will start to believe that you're performing brain uploading, and people will want to protect these simulations as people with rights, or ban it to prevent conscious mini-mes from being booted into a simulation and psychologically tortured like the guy in Altered Carbon. Then of course theere's the suicide cults like in the Cowboy Bebop episode "Brain Scratch" - leave your aches and pains behind !

I pray it doesn't get that weird! But if I had kids I wouldn't be raising them to talk to ghosts!

[ Parent | on HN ]

Re:A simulation of me: fine-tuning an LLM on 240k text messages by pixl97 on Thursday January 04, 2024 @08:51PM (#38874768)

> Maybe some extremists will think its unholy

At the other side of this equation is the extremists that think this is great. To go full Godwin's law there are certainly people out there that would like for a digital Hitler to stick around forever.

[ Parent | on HN ]

Re:A simulation of me: fine-tuning an LLM on 240k text messages by HKH2 on Thursday January 04, 2024 @11:11PM (#38875608)

I'm sure people will create holographic Hitlers for historical reasons too.

[ Parent | on HN ]

Re:A simulation of me: fine-tuning an LLM on 240k text messages by prepend on Friday January 05, 2024 @08:05AM (#38878649)

> I also get the vibe that digital cloning will be popular.

I think it will end up just being boring. The sad reality is that no one really wants to talk to us, except for a few who are amazing poets or something.

I think what’s more likely is to train models on all material created and ask useful questions for purposes of probate (“when did you buy the house”) because its faster and cheaper than manual research. Or maybe for family history summaries (“how many cousins lived in Colorado in the 1990s and moved to Iowa?”)

[ Parent | on HN ]

Re:A simulation of me: fine-tuning an LLM on 240k text messages by jononor on Friday January 05, 2024 @12:17PM (#38881646)

People have a strong need to feel connected to others. Many feel particularly strong about their ancestors, or perhaps more distant family or people that were there in particular moments in life. I can definitely see a possibility that people talking to AI emulations of their grandma/godfather/mentor/etc from time to time (after that person is dead). It will of course depend on how culture evolves, maybe it has little place in a hyperinvidualistic world view.

[ Parent | on HN ]