The Hitchhiker's Guide to Monoculture

2026-07-12

AI Summary

This study investigates whether AI-powered coding assistants lead to homogenization of developer code at both syntactic and semantic levels. Leveraging Kaggle competition submissions from 2019 to 2026, the work introduces a dual-perspective analysis—combining TF-IDF for syntactic similarity and Voyage 3 code embeddings for semantic distance—to systematically quantify shifts in code diversity at both submission and competition granularities. The findings reveal that AI assistance significantly increases syntactic similarity, manifesting as convergence in code structure and reduced variation across syntactic dimensions. In contrast, semantic distances remain stable over time, indicating that the diversity of problem-solving strategies is largely unaffected. This research provides empirical evidence on the nuanced impact of large language models on software development ecosystems, highlighting a divergence between syntactic conformity and semantic resilience.

Abstract

Large language models (LLMs) often produce homogeneous outputs, raising concerns that AI coding assistants may lead to convergence in the software artifacts that developers create. Whether this occurs in practice is unclear because developers interactively prompt, evaluate, modify, and reject model outputs, and because outputs vary with prompt and repository context. I examine code homogenization using Kaggle contest submissions from 2019 to mid-2026. I first document widespread convergence toward the random seed value 42, consistent with LLMs reinforcing a longstanding convention in programming culture. I then study homogenization more broadly, at two levels of aggregation and abstraction. At the submission level, I measure the average pairwise similarity of submissions within contests. At the contest level, I measure the conceptual span of submitted code, motivating distinct measures for each: TF-IDF representations, which capture surface syntax, and Voyage 3 code embeddings, which capture code intent and semantics. The results demonstrate substantial syntactic homogenization at both the individual and collective levels: individual submissions have become more alike in literal syntax and code structure, while the latent dimensionality of syntactic variation has narrowed. In contrast, I find little evidence of semantic homogenization, individually and collectively. Average semantic distance remains essentially flat, and the contest-level latent dimensional span of semantic approaches remains stable, with evidence suggesting it has even expanded modestly. These findings suggest that AI coding assistants are certainly standardizing implementation details, yet they have not yet produced evidence of homogenization in the approaches and problem-solving strategies coders employ.

Problem

• code homogenization
• large language models
• AI coding assistants
• semantic diversity
• syntactic convergence

Innovation

• code homogenization
• large language models
• semantic embeddings
• TF-IDF
• Voyage 3