{"slug": "artificial", "title": "Artificial", "summary": "Mimi and Chee discuss the design choices behind large language models, arguing that LLMs do not need to generate human-like language and that their personalities are a training choice, not a necessity. They explore alternative uses for LLMs, such as generating code for tools like Patchwork, and critique the current trend of making models sycophantic and expensive.", "body_md": "*it*works with the Humans\"\n\n**mimi**[submit a](https://feelingof.london/demo)demo\n\n*consciously*making a little joke when they say it?\n\n*Claude*is not a person. Right?\n\n*really*don't like it, but i don't know why yet\n\n**mimi**\n\n**mimi**`claude`\n\nprogram does this automatically when you use it to commit code. and i think github when you use copilot?\n\n**mimi**`—`\n\n, `;`\n\n, proper capitalization) also make me wary now*seem right*at first blush, but have subtle mistakes the kind a person never makes\n\n*need*to write something\n\n**chee**\n\n**mimi**[felienne](https://www.youtube.com/watch?v=-Br66SUjsdQ&t=12129s)hermans\n\n*make*mistakes\n\n**mimi*** have*to be assistants, they don't have to talk in prose\n\n[Natalia](https://uxdesign.cc/when-ai-passes-the-capitalist-turing-test-18baacbcf18f)Talmina came in to talk about ai and conversation she mentioned that there's linguistics research that people adapt to each other's mental states when they are talking to each other\n\n**mimi**\n\n**chee****from people** outside of https://claude.ai and vscode\n\n*part*of why the models are SO compute expensive and therefore money expensive. The output of pre-training is a base-model.\n\n*fine-tune*the model by feeding it loads of conversations in the style that we're aiming for (sycophantic simpering) and retrain some layers of the model towards the target output.\n\n*instruct*models rather than\n\n*base*models.\n\n```\n{\n    \"user_message\": \"Generate a tool that does X\",\n    \"model_response\": \" { @patchwork: more javascript json stuff goes here, blah blah blah you get the idea}\",\n    ...\n}\n```\n\nAnd then it would, hopefully, be really good at generating the javascript for patchwork tools.\n\n**mimi*** LLMs don't have to generate human-like language AT ALL.*It is a design / training choice that LLMs generally 'have' personalities. We could train an LLM to generate Patchwork tools or JavaScript and they wouldn't have any capabilities to generate human-like language like most of them do now. Our input could\n\n*also*be entirely different / richer than a text description of the thing we want out.\n\n*exactly*what i want\n\n[We've had a go at another](https://www.inkandswitch.com/patchwork/notebook/chitter-chatter/)way , but there are so many more to try :))\n\n*great*. do we wanna talk about the stuff re: a new punk? punk computing? like, about craft vs art that's veneer and music with guttural screams and mistakes left in and... etc?", "url": "https://wpnews.pro/news/artificial", "canonical_source": "https://www.inkandswitch.com/tangents/artificial/", "published_at": "2026-06-21 13:22:47+00:00", "updated_at": "2026-06-21 14:05:27.057064+00:00", "lang": "en", "topics": ["large-language-models", "ai-products", "ai-tools", "ai-research", "ai-ethics"], "entities": ["Claude", "GitHub Copilot", "Felienne Hermans", "Natalia Talmina", "Patchwork", "Ink & Switch"], "alternates": {"html": "https://wpnews.pro/news/artificial", "markdown": "https://wpnews.pro/news/artificial.md", "text": "https://wpnews.pro/news/artificial.txt", "jsonld": "https://wpnews.pro/news/artificial.jsonld"}}