← Back to articles
Researcher attempts to fine-tune a Chinese LLM to perfectly regenerate Borges' 'Pierre Menard' story token-by-token rather than merely imitate it.
LessWrong AI · April 20, 2026
AI Summary
- •The goal is not to create a Borgesian-style story but to have the model generate the exact text of 'Pierre Menard, Author of the Quixote' through its weights
- •Initial approach involved using Kimi K2.5-Thinking with Borges' complete biography and literary influences in the system prompt, but the 256K-token context window proved insufficient by five orders of magnitude
- •The project treats this as a technical challenge distinct from simple memorization or machine transcription, requiring the model to generate text that perfectly coincides with the original work
Related Articles
Large Language Models
Skild AI acquires Zebra Technologies' robotics business, gains software that controls any robot without knowing its design
Robotics & Automation News·Apr 21, 2026
Large Language ModelsOpen-Source AI
Openheim launches open-source AI agent in Rust — developers can now run autonomous AI tools locally without cloud dependency
Hacker News·Apr 21, 2026
Large Language Models
1Password used AI agents to break apart its monolithic codebase — and found the approach faster and cheaper than human refactoring
Hacker News·Apr 21, 2026
Large Language Models
Salesforce launches Headless 360, opening its entire customer-data platform as a backbone for AI agents to automate business tasks
Hacker News·Apr 21, 2026
Large Language Models
Researcher Giles Thomas shares improved instruction fine-tuning results for custom AI language models, showing concrete performance gains in training efficiency
Hacker News·Apr 21, 2026
Stay ahead with AI news
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started Free