Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

edwarddonner.com

Post reply on HN
Print view View original

21–30 of 145 posts

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

#21

Post by jazzyjackson »

alchemist1e9 wrote: post #13
mewpmewp2 wrote: post #8

Earlier quoted context omitted.

> it might be therapeutic to have a conversation with a deceased loved one. Would it? I'm not a specialist, but it would seem a bit off to me to approach it like that. I'd think best is to accept that your deceased one has gone, and it seems like having a fake one might just prolong the grievance process and possibly also develop into an obsession of trying to replace your deceased one with the simulation, upgrading…

Have you ever watched the original Superman movies and Clark talks with the AI hologram of his biological father from his home planet? I believe in the future our descendants will likely find it normal to talk with AI versions of their ancestors, perhaps not us parents right now, but I suspect my children will willingly train and generate AI versions of themselves for future generations to interact with.

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!

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

#22

Post by lomir8 »

AriedK wrote: post #10
wyldfire wrote: post #2

> I immediately recognized this would be in extremely poor taste and removed the data. Enough on that topic! Poor taste for a blog post. But I wonder if LLMs could advance to a point where it might be therapeutic to have a conversation with a deceased loved one.

It has been tried. Joshua Barneau tried it with GPT-3, training it on his deceased fiancees text- and facebook messages. It’s a moving read: https://www.sfchronicle.com/projects/2021/jessica-simulation... It ‘worked’ for him for a while.

Shocking a biochemical reaction ceases when external stimuli is removed.

Technology is like acupuncture and astrology of the modern west.

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

#23

Post by vunderba »

I’m far from the first to think of this. Several people — perhaps inspired by creepy Black Mirror episodes — have tried to fine-tune an LLM on their SMS or WhatsApp history in an effort to create a simulation of themselves.

It's a much older concept than Black Mirror. Ever since Markov chain IRC bots got popularized in the late 90s and early 2000s, people have been trying to train their virtual doppelgängers. I'm sure it won't be long before you can pay to have a post-Mortem ghost of yourself duplicated in a cemetery seance style MMO to perpetually live on and interact with other people who have passed on.

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

#24

Post by 0xEF »

AndrewKemendo wrote: post #20

I’m glad someone put the time it to show it can finally work. Would love to do this too, but have a gazillion other projects that are priority.

I also considered it, but the existential dread of discovering myself to be an insufferable numpty pretty much quelled that fire.

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

#25

Post by pixl97 »

jazzyjackson wrote: post #21
alchemist1e9 wrote: post #13

Earlier quoted context omitted.

Have you ever watched the original Superman movies and Clark talks with the AI hologram of his biological father from his home planet? I believe in the future our descendants will likely find it normal to talk with AI versions of their ancestors, perhaps not us parents right now, but I suspect my children will willingly train and generate AI versions of themselves for future generations to interact with.

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…

> 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.

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

#26

Post by HPMOR »

vunderba wrote: post #23

I’m far from the first to think of this. Several people — perhaps inspired by creepy Black Mirror episodes — have tried to fine-tune an LLM on their SMS or WhatsApp history in an effort to create a simulation of themselves. It's a much older concept than Black Mirror. Ever since Markov chain IRC bots got popularized in the late 90s and early 2000s, people have been trying to train their virtual doppelgängers. I'm sur…

This actually seems like a really good startup idea.

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

#27

Post by teitoklien »

vunderba wrote: post #23

I’m far from the first to think of this. Several people — perhaps inspired by creepy Black Mirror episodes — have tried to fine-tune an LLM on their SMS or WhatsApp history in an effort to create a simulation of themselves. It's a much older concept than Black Mirror. Ever since Markov chain IRC bots got popularized in the late 90s and early 2000s, people have been trying to train their virtual doppelgängers. I'm sur…

> Pay to have a ghost of yourself duplicated in a cemetery.

I mean, they already do it at funerals [1]. I don’t see why not, would make cemetery a fun hangout place to talk to people from history.

[1](https://m.youtube.com/watch?v=s2FJRbRsBBY)

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

#28

Post by sir0010010 »

wyldfire wrote: post #2

> I immediately recognized this would be in extremely poor taste and removed the data. Enough on that topic! Poor taste for a blog post. But I wonder if LLMs could advance to a point where it might be therapeutic to have a conversation with a deceased loved one.

You can already do this without LLMs: give someone else all of your loved one's writing samples and ask them to do their best to text you in a style as if it was the loved one.

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

#29

Post by xbmcuser »

mewpmewp2 wrote: post #8
wyldfire wrote: post #2

> I immediately recognized this would be in extremely poor taste and removed the data. Enough on that topic! Poor taste for a blog post. But I wonder if LLMs could advance to a point where it might be therapeutic to have a conversation with a deceased loved one.

> it might be therapeutic to have a conversation with a deceased loved one. Would it? I'm not a specialist, but it would seem a bit off to me to approach it like that. I'd think best is to accept that your deceased one has gone, and it seems like having a fake one might just prolong the grievance process and possibly also develop into an obsession of trying to replace your deceased one with the simulation, upgrading…

Yours might be a western view point in most Asian cultures even many African cultures they view death and ancestors differently than the way people do in the west. So such a service might get quite popular

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

#30

Post by krainboltgreene »

HPMOR wrote: post #26
vunderba wrote: post #23

I’m far from the first to think of this. Several people — perhaps inspired by creepy Black Mirror episodes — have tried to fine-tune an LLM on their SMS or WhatsApp history in an effort to create a simulation of themselves. It's a much older concept than Black Mirror. Ever since Markov chain IRC bots got popularized in the late 90s and early 2000s, people have been trying to train their virtual doppelgängers. I'm sur…

This actually seems like a really good startup idea.

This is definitely what silicon valley would consider a good startup idea.
Post reply on HN