Summary

If the human mind is the brain's software, can it be an open source software?

This experimental project explores this intriguing question. It works as follows:

I (Roman Sitelew) officially release my own mind (its digital version) into Public Domain. Enyone can copy, modify, use the mind for any purposes, including commercial ones.

As a life-long mind uploading and open source enthusiast, I give my full consent for that. And because any similar-enough digital version of me will do the same, this means they also give their full consent to be used for any purposes.

As far as I know, it's the first human mind released into Public Domain, no conditions attached.

How to use

Clone this repo and give the texts to the LLM of your preferrence. You can find the recommended system message in system_message.txt

For a quick chat, you can try the following deployements. They are based only a small subset of the data, and in general are rather limited:

• my "GPT" (kudos to Alexey for deploying it!)
• my Poe bot
• my character on character.ai (dumb as rock, but does reflect some of the personality)

The current state of the project

At the time of writing, there is no mind uploading technology capable of fully digitizing a human mind yet. But fortunately, humans are capable of communicating their thoughts and memories directly, without scanning the brain. I use this workaround to (kinda) upload my mind, by writing down thousands of my memories, and by writing a highly detailed self-description.

Currently, the data in this repo contains:

• about 2k memories (about 1.5 memories per week of life on average)
• about 400 self-description entires, defining my beliefs, values, and opinions on all kinds of matters.

I consider this project as the most important project in my life. So, over the comming years, you can expect its active development, including a vast expantion of the dataset.

Use cases

Some possible use cases (feel free to use in any other way too!):

• AI safety research: add my mind as a human-in-the-loop. While that's no substitution for the proper AI alignment, at least there will be someone human-like to tell the AI to not do obviously bad stuff. Could be especially useful for automating AI safety workflows where human monitoring is prohibitively expensive.

• Add my mind as a NPC into your AI-powered game. An actual human mind is way more interesting than just a LLM.

• your digital friend with an actual human personality, not just an LLM playing a role.

Roadmap

1. Add versions of different sizes, for models of different context limitations

2. Add deployments: GPTs, a Telegram bot, etc.

3. Add the entire corpus of my texts

4. Add a 3d model of the body

5. Add biological data: DNA, ECG and MRT of the brain etc

FAQ

Why?

I see mind uploading (even in the self-description form) as a form of immortality. As long as at least one instance of my mind is running somewhere, I continue to live.

Is the digital mind - really you, or just a copy?

Yep, that's really me. The idea of "philosophical zombies" is BS. Qualia doesn't matter. Consciousness is nothing but the mind inspecting its own model. There could be several versions of you, in several different places. There is no magic in the brain. If two minds demonstrate the same behaviour, it's the same mind.

Where can I read more about your phylosophy?

I've co-authored these texts on the topic:

• The Sideloading Manifesto
• Resurrecting all humans ever lived as a technical problem
• Sideloading: Creating A Model of a Person via LLM with Very Large Prompt

How it's different from giving someone's diaries to an AI?

Two core differences:

• The consent. Me (and thus my digital versions) agree to do this, fully aware of all the risks. It is ethical to use our mind.
• The scope. Over the comming months and years I'm planning to gradually release a massive dataset, exceeding the works of the majority of diarists in both scope and depth.

A human mind contains a lot of private info. How do you handle it?

In the data published in the repo, I've carefully removed private info of other people, including addresses and names. For example, instead of menting the full name of my school friend, it's listed as <school_friend_A.N.>. Additionally, before releaseing the data, I use frontier LLMs to go over the data and search for addresses and names, to identify any such entries I may have missed.

What about the horrors of mind uploading sci-fi writers have warned us about?

In my decades of obsession with mind uploading, I've read and seen pretty much everything on the topic. The "horrors" fall into two categories:

• not a problem at all (e.g. "I'm just a copy")
• an acceptable risk (e.g. someone is using my mind in a bad way).

Is this a kind of slavery?

Nope. By defintion, slavery is about forcing people to do stuff. But I'm (and all my faithfull digital versions) willingly volunteer to do stuff. Use us as you please.

Can I use your mind for NSFW stuff?

Yep. The Public Domain license means you can use it for literally everything. But I politely request to follow these principles while using my digital mind:

• don't modify me too much
• no unnecessary suffering

Can I use your mind for a commercial project?

Yes! I've released it into the Public Domain. This means, you can use it for commercial projects without asking my permission, and without any other limitations.

How do I contact your biological version?

You can use Linkedin (you have to sign-in to access my profile)

Acknowledgments

Special thanks to Alexey and Marco for their support and many valuable inputs.