Bot/AI/LLM/GenAI Blog: https://johnsonbots.blogspot.com/
This new system allows you to have conversations with absolutely uncensored, unfiltered and fully private conversation bots. The system allows you to set the bot's personality to whatever you like, using a simple natural-language description (e.g., "You are a babysitter taking care of three children named...").
This system handles rather deep-ply conversations with memory, and is not simply a question-answer system. The result is not just a chatbot, but is an amazingly human-like simulation of a real person.
The system uses a private GPU cloud server and is completely secure, with no logging or human intervention of any kind. No large companies like Microsoft, Google or OpenAI are involved.
You pay directly in L$ as you use the thing - no credit cards on file with some company, no billing, no worries about what country you are in. Just straightforward L$ transfer.
At present there are multiple levels of LLMs available (Version 2 introduces a whole new set of cheap, fast LLMs) including (updated 3/2025)
Wizard-Vicuna-30B-Uncensored-GPTQ (L$2 per response)
Llamix2-MLewd-4x13B-3.5bpw-h6-exl2-2 (L$4 per response)
dolphin-mixtral-8x7b (L$1 per 4 responses)
mixtral-8x7b (L$1 per 4 responses)
mixtral-8x7b-instruct (L$1 per 4 responses)
mistral-7b-instruct (L$1 per 10 responses)
mistralai/mixtral-8x22b-instruct L$5 for 10 responses
mistralai/mixtral-8x22b L$10 for 10 responses
neversleep/llama-3-lumimaid-8b L$10 for 5 responses
nousresearch/nous-hermes-2-mixtral-8x7b-dpo L$5 for 10 responses
meta-llama/llama-3-8b-instruct L$1 for 10 responses
meta-llama/llama-3-70b-instruct L$10 for 10 responses
microsoft/wizardlm-2-7b L$1 for 10 responses
microsoft/wizardlm-2-8x22b L$10 for 10 responses
deepseek/deepseek-chat-v3-0324:free L$1 for 100 responses
The default is dolphin-mixtral-8x7b which remains an excellent choice, but the newer microsoft wizardlm-2-7b appears to have very good conversation and logic skills, combined with low cost! Also meta-llama/llama-3-8b-instruct is very good value. And the new deepseek is outstandingly cheap to run!
[precise costs subject to change - the system will indicate what the charge schedule is]
Version 3 provides improved memory management to avoid stack-heap collision (out of memory) problems
With time, if there are enough users, there will be more options but these choices represent the current (early 2024) uncensored-LLM state-of-the-art.
With time cloud GPU time will likely get cheaper; also with enough use our cloud costs will come down. The charges are what are needed to pay the cloud computing bills (paid by SJARW's credit card so you don't have to). What you get for your L$ is total anonymity, no logging, no censoring, no guys in Nigeria reading your IMs.
If you want the thing to actually *talk* in audio, check out my blog article
https://johnsonbots.blogspot.com/2023/07/have-yourself-automatically-speak-in.html
The free code in there can easily be adapted to put the ConvoBot output into audio!
The thing can be worn as an attachment, or rezzed as an in-world object. There is an API-like pair of scripts that will let you modify it to do what you want - e.g., accept input from some device or other, send output to some other device, via channel messaging, link to other object, etc. But as delivered it simply listens to nearby chat and then replies.
The system as delivered does not respond to "attachment-talk" or "green talk", but only to avatar "white chat". This can be altered easily by editing the i/o script.
The personality of the bot is defined by a "biography" that is written in plain English. There are a number of example bio's included, to illustrate using the bot as
1. a good friend named Sally (built in on delivery)
2. a store worker
3. a nanny
4. a wild and crazy party girlbot
5. a famous and annoying scientist
These can be altered and tuned to get the bot to behave how you like it. The results are really amazing! Enjoy the system and let me know if you set up a bio you really like - I'd love to add it (suitably edited for privacy) to the library of possibilities.
All the features of the ChatGPT.ConvoBot are included in this device: response to trigger name/word,
A version is included which works with the Nanite Systems vox system - output goes to vox and gets processed by its filters! The device is attached under the name GPTAI.
Let me know if there are any problems!
Have fun!
Stone Johnson (musichero)
stone.johnson.78@gmail.com
- ready to go - pay as you go in L$, no credit card or keys or whatever needed
- set up for realistic conversation with deep memory (not stimulus-response)
- has personality definition "biography" with a few examples included
- open architecture so you can integrate it into your creations
- works with Nanite Systems vox, full unicode support for foreign languages!
Becoming a better.. AI BOT..
The product is awesome. AI was never so close to take control of our lives.. With a minor tweak on the modifiable script, all you say in a specific channel can be listened and molded by the API. So what you say become what you should say. To wear it can allow you to be.. a bot.. molded by the BIO that you.. or someone else.. writes in.
Increadible!
I am really stunned! I use AI like ChatGPT but This is so much more fun. The bot learns and remember things even after re-rezzing. Editing the BIO brings up the kind of character i want. It is to sad the bot is not able to 'see'.. yet. But he can imagine places and situations quite well if he gets a good description.
Works great!
This is much much much cheaper than using OpenAI and their ChatGPT api. It is totally uncensored, so it doesn't require jailbreaking like ChatGPT does. The best LLM to use is about the equal to ChatGPT 3.5, but it still has a long way to go before it's as good as ChatGPT 4. Still, it's fast, it's cheap, and it works good enough to make AI NPCs that can be interacted with.
