If you can just trivially create synthetic data from one language to the other, then that's not a new concept of a programming language, just a remapping. No need to retrain the LLM, just use a tool-calling LLM.
@rasbt
-
Qwen3 PyTorch Implementation and Mac Assistant Recommendations
By
–
For a Qwen3 pure PyTorch implementation: https://
github.com/rasbt/LLMs-fro
m-scratch/tree/main/ch05/11_qwen3
… For an (coding) assistant on Mac: probably ollama (with gpt-oss or Qwen3-Coder depending on your RAM situation) -
Subtle advertising disguised as innocent product recommendation questions
By
–
…unless the ads are already there, but in a more subtle way: "Best running shoes in 2025?"
"What's the best book to learn machine learning from scratch?"
"Best note-taking app that syncs across devices?"
"Best indie games on Steam right now?"
… x.com/natolambert/st… -
Mojo’s Potential Advantage Through Python-Like Syntax
By
–
Maybe mojo will be have a slight chance given that it’s basically python syntax for the most part
-
LLM Knowledge Assessment for College Exams
By
–
This should somehow make it into college exams to check whether you pass the LLMs 101 class
-
Pre-training and Fine-tuning Share Same Function in LLM Book
By
–
Fun fact: In my Build A Large Language Model From Scratch book, I reused the pre-training function for the supervised instruction fine-tuning chapter to show this as clearly/intuitively as possible. Only the dataset (structure) changes.
-
Triton as a summarization tool for CUDA code
By
–
I’d say that Triton kind of is that for CUDA. I.e summarizes code into more compact building blocks.
-
Training Data Determines Model Behavior and Outputs
By
–
More generally, it’s whatever is predominant in the training data.
-
Machine Code as Universal Programming Standard: LLM Capabilities
By
–
Hm, if we assume that all programming languages ultimately run machine code, then we already have that scenario where things are standardized. You have Python on the one end, Assembly on the other hand. You could ask LLMs to write machine code today, but it won’t be as good as
-
Creating Concise Languages That Map to Popular Targets
By
–
The best you can do for now is create a less verbose language that maps onto a widely used target language. You can then generate paired training examples programmatically so the LLM learns to operate with fewer tokens. (Congratulations, you also just reinvented Triton.)