Andrej Karpathy announces the release of GPT-4.5 by OpenAI, sharing his enthusiasm for a model he has been anticipating for roughly two years since GPT-4's launch. He highlights that this release provides a qualitative measurement of the slope of improvement achievable through scaling pretraining, suggesting it serves as a key benchmark for understanding how much progress comes from simply scaling up model training. The post includes an interactive comparison, likely contrasting GPT-4.5's outputs against previous models to illustrate these gains.

Andrej Karpathy shares his evolving approach to using LLMs for building personal knowledge bases across various research topics. He notes that a significant portion of his recent token usage has shifted away from code manipulation and toward organizing and structuring knowledge.

This reflects a broader trend among AI researchers and engineers who are finding that large language models serve not just as coding assistants but as powerful tools for knowledge management and research synthesis. Karpathy's insight highlights how LLMs are increasingly becoming integral to intellectual workflows beyond software development.