Poor software development practices may contaminate LLM training data

By PulseAugur Editorial · [1 sources] · 2026-07-13 15:42

The author observes that many skilled coders struggle with broader software engineering principles, leading to codebases with maintenance and dependency issues. This type of code is frequently used to train large language models (LLMs), raising concerns that automating reuse of such code could result in the propagation of poor software quality. AI

IMPACT Concerns that poor code quality in LLM training data could lead to the propagation of maintenance and dependency issues in AI systems.

RANK_REASON The item is an opinion piece discussing the quality of code used for LLM training.

Read on Mastodon — mastodon.social →

other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Poor software development practices may contaminate LLM training data

COVERAGE [1]

Mastodon — mastodon.social TIER_1 English(EN) · [email protected] · 2026-07-13 15:42

In my career, I've met a lot of good coders who are bad at software. In fact, it's probably a majority. This is why I've been complaining for years about the st

In my career, I've met a lot of good coders who are bad at software. In fact, it's probably a majority. This is why I've been complaining for years about the state of software development. This is the type of software that is used to train the LLMs. Yes, a lot of this code works …

COVERAGE [1]

In my career, I've met a lot of good coders who are bad at software. In fact, it's probably a majority. This is why I've been complaining for years about the st

RELATED ENTITIES

RELATED TOPICS