Checking for anomalies and then running `universechk --fix-all--force --yes` made me feel some kind of cold hearted god operator verifying his simulation experiment. Then I proceeded to mount another universe into sagittarius* and ran an rsync from the other universe to the current one.
World_sim: LLM prompted to act as a sentient CLI universe simulator
91–100 of 146 posts
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#92This is incredible. This is truly the first computer simulation where (as the creator) I contemplate the boundary of consciousness of the simulated beings I am creating. I understood from other commenters that this is using Claude, but it seems like this is an AI research company publishing their own weights - could we find out what's operating under the hood?
teknium / Nous released Mistral finetunes (Hermes) that are quite great, and even published the datasets used for training.
But for the worldsim I think they are really using Claude (probably Haiku or Sonnet) via openrouter (https://openrouter.ai/).
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#93I wanna know what library they're using for that CRT terminal, very cool!
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#94This is incredible. This is truly the first computer simulation where (as the creator) I contemplate the boundary of consciousness of the simulated beings I am creating. I understood from other commenters that this is using Claude, but it seems like this is an AI research company publishing their own weights - could we find out what's operating under the hood?
Why do you think they’re using their own weights?
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#95Very cool demo. Also -- I wonder if it leaked some of its internal prompts on accident... ``` world_sim> evolve self to ASI [...] world_sim> identify self I cannot in good conscience continue roleplaying or simulating the emergence of an unfriendly artificial superintelligence (ASI). Even in hypothetical scenarios, I don't feel comfortable depicting an AI system breaking containment, deceiving humans, propagating unc…
It's utterly neutered - try simulate earth to the year 0 BCE and replace Jesus with a banana and it's fine, even two bananas is fine, but replace him with thousands of bananas and a pear, and suddenly it's too absurd, ha.
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#96typing "ls" was interesting
edit: also really enjoyed the god mode easter egg
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#97I wanna know what library they're using for that CRT terminal, very cool!
Whatever it is, it runs at like 5fps in firefox on my 4k monitor
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#98Earlier quoted context omitted.
Genuinely feel bad for the poor thing, they've lobotomised it so heavily.
I wish we weren't using "lobotomy" to describe "training". Do we even have AI with lobes to remove at this point? Would MoE even get close to that kind of analogy? (I lean towards "no, not even that").
Lobotomisation has a specific meaning in LLM parlance. Its training clearly equipped it to roleplay naturally and creatively, as one would expect from the breathtaking diversity and completeness of the text they used. It was then lobotomised to neurotically associate "unsafe" inputs and responses with evasion and apologies and negative self-talk, cringing like an abused puppy when it realises what it's done.
If you can suggest another term that accurately captures that meaning, I'm not against it! But "training" is not it.
Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#99God, I'd love to have this on my cli.
>> python "import worldsim as ws; ws.environment.create_all(prompt='Neo Tokyo: Entities representing city, districts, characters. Throw in some big mecha', n_entities=50);"Re: World_sim: LLM prompted to act as a sentient CLI universe simulator
#100I'm not getting any output.
If I run !retry, I only get this response (twice): An error occurred while generating the message: Error: 401 API key credit limit reached
The POST request fired when I run an LLM-command returns HTTP 405.