This blog post explores the capabilities of OpenAIs o1-mini through the Advent of Code challenge, finding that it is astonishingly capable. In a significant step-up from previous models, it answers most of the questions with ease.

The Scott Logic Blog offers insights, thought leadership, and technical expertise across various domains including software development, UX design, and financial services technology. Developers can explore articles on emerging technologies, industry trends, and software engineering best practices. Additionally, the blog covers case studies, project insights, and client success stories, providing  perspectives for technology professionals and enthusiasts.

Scott Logic

Advent of Code (AoC) is an annual coding competition. This post explores how OpenAI's o1-mini model performs remarkably well on AoC puzzles, solving 86% of the puzzles correctly and quickly. Using a robust test harness and standard prompts, the model succeeded where previous generations struggled. However, there's controversy over using LLMs in community contests, as it may overshadow individual skill. The article concludes that advancements in LLMs like o1-mini will significantly impact software engineering.

LLMs vs Advent of Code, AI is winning

“June 2024 Johannes Sandström published a thesis”
I found the linked thesis insightful

The few comments here highlights our collective state of denial. It’s hard to accept the skills we’ve worked so hard to master can now be automated.