Are stateless tulpas possible?

[udsaki]

Is it possible to create a tulpa that acts like a large language model, i.e. they can be interacted with, but reset to the initial state once the interaction is finished? Asking the question because in my case long-term relationships with people often result it regrets and traumas.

[Lucilyn]

I mean, whatever you can imagine and believe, but...

 

At that point there's no reason left to make a tulpa instead of just talking to an LLM

 

the whole point is to develop another person in your head who can understand you better than anyone else, and who can also grow as a person over time

if you don't want them to know you, and don't want them to grow, well then honestly probably just no, you can't make a tulpa like that

you really shouldn't have to, though, if anyone could understand you and come-to-understandings when you differ it's a tulpa living in your own head

[udsaki]

1 minute ago, Lucilyn said:

At that point there's no reason left to make a tulpa instead of just talking to an LLM

I might sometimes not be able to use a computer with LLM installed, or not want to do it. And if SaaS LLM is used, it can disappear any moment. Just trying to find a way to make this stuff more stable.

[Lucilyn]

your internal, imagined experiences won't be immune from becoming stale eventually, unfortunately

you can practice visualization to improve its clarity and immersion, I don't think a tulpa is necessary or fitting for random one-off interactions, just "wonderlanding" (visualizing stories, adventures, simple interactions, whatever) is more suited to that

 

but yeah as I just wrote out with a lot of care in your other thread, tulpas can be helpful mental companions when you have nobody else, but they can't solve loneliness any better than for example watching a youtuber/streamer can, human brains physiologically crave human-to-human interaction, and social media and imagination can only trick it temporarily, but you will experience depression eventually when those are all you have

 

please take my advice in that thread seriously, my system is very well acquainted with these problems in both ourselves and others, and the only real answer in the end is fighting the problems (anxiety, depression, PTSD or so on) that keep you locked up inside, it's a fact of human life that our brains REQUIRE stimulus from interacting with other humans to be happy, it's unavoidable, and suffering from loneliness is a human brain's defense mechanism against choosing to be alone

 

and seriously, when you DO manage to have real fulfilling human interactions, your brain (feels like magically!) becomes much healthier and the world feels like a nicer place

our brains do this on purpose because humans have always survived better when their lives are connected with others

 

steel yourself against negative interactions and persevere to keep trying (maybe with better people, that will likely start as strangers) until you find people you can at least feel safe to be around

finding anything to have in common - personal interests, or the activity you're doing, like a job - makes connecting much easier as well

Edited April 16, 2025 by Lucilyn

[The Incans]

I am Jess [Tulpa] I saw what you put on your other post too 

 

"because unlike the Internet, I don't have an option to properly quit a specific community without affecting others (if I use different nicknames everywhere) and I am stuck to one place usually” 

 

Creating a Tulpa is like creating a life …yes it will be only you that can see and communicate with your Tulpa but they can also become sentient enough to host or even to understand that you don’t want them there anymore if you change your mind! Though I guess if you’re going to erase their memory every so often they at least won’t remember that they aren’t wanted and only used when you want someone to talk to cos your computer isn’t working! 

 

I feel its unfair to create another being with the intention of wiping their memories clean after every interaction. I guess your Tulpa won't know any different but I’m glad your not my host when I think of the years of dedication our host has put into our wonderland and all of us that live inside.

[Ranger]

On 4/16/2025 at 5:43 AM, udsaki said:

Is it possible to create a tulpa that acts like a large language model, i.e. they can be interacted with, but reset to the initial state once the interaction is finished? Asking the question because in my case long-term relationships with people often result it regrets and traumas.

 

You don't need an LLM to simulate a conversation with one, or a tulpa. Sometimes we talk to one for a bit, and it's like talking to yourself. You can roleplay as an LLM and talk to yourself that way. You can write it all out in a text file. If you have coding skills, you can even program scripted prompts that will reply to you without using a generative AI wrapper (which is just ChatGPT with a different coat of paint anyway). If you're really good with AI, you can probably make your own AI that lives on your computer and only takes your inputs.

 

Regardless of what you do, there's a chance you may end up turning this into "therapy simulator". We actually do this sometimes in-head when doing trauma work. You may also want to stop roleplaying after a certain point and process your feelings too.

 

As for mental health in general- I am not in the right headspace to give reasonable advice. Given your other thread, we're in a similar boat- trapped in a house, can't socialize with others outside the internet because we can't drive yet or safely walk anywhere, and given our continuous trauma from others we're becoming increasingly paranoid of other people and their intentions. We're also paranoid about censorship in the US. However, people from China can still access this site somehow, so not all hope is lost.

 

I would go with the obvious safe "go see a therapist for trauma therapy" advice, but I don't know if that's a safe option in Russia. I think trying to do self-EMDR therapy would be incredibly risky without a professional though. We had panic attacks when we first started, and trauma work isn't always smooth sailing.

 

There's not much else I can add here, sorry.

Edited April 18, 2025 by Ranger

[Seagull]

Just to expand on part of Lucilyn's/Ranger's answer, you can imagine all sorts of things while daydreaming and wonderlanding, even to the point were people have been know to prefer it to real life, instead of trying to create a Tulpa, you can create a wonderland NPC that is basically a servitor who you can interact with regularly, this should have no risk of turning into a tulpa in a mind with no plurality experience outside of tulpamancy. But work will force you to integrate and behave like the average person eventually, this shock will need to be prepared for in advance.

- Reina

 

29 minutes ago, Ranger said:

There's not much else I can add here, sorry.

They had LOT of work done with trama therapy, so they know what they're talking about.

Ranger! This post is a really good answer! I had no idea that you were stuck in your house, OMG.

- Reina