Developer Challenges on Large Language Models: A Study of Stack Overflow and OpenAI Developer Forum Posts
Bibliographic record
Abstract
Large Language Models (LLMs) have gained widespread popularity due to their exceptional capabilities across various domains, including chatbots, healthcare, education, content generation, and automated support systems. However, developers encounter numerous challenges when implementing, fine-tuning, and integrating these models into real-world applications. This study investigates LLM developers' challenges by analyzing community interactions on Stack Overflow and OpenAI Developer Forum, employing BERTopic modeling to identify and categorize developer discussions. Our analysis yields nine challenges on Stack Overflow (e.g., LLM Ecosystem and Challenges, API Usage, LLM Training with Frameworks) and 17 on the OpenAI Developer Forum (e.g., API Usage and Error Handling, Fine-Tuning and Dataset Management). Results indicate that developers frequently turn to Stack Overflow for implementation guidance, while OpenAI's forum focuses on troubleshooting. Notably, API and functionality issues dominate discussions on the OpenAI forum, with many posts requiring multiple responses, reflecting the complexity of LLM-related problems. We find that LLM-related queries often exhibit great difficulty, with a substantial percentage of unresolved posts (e.g., 79.03\% on Stack Overflow) and prolonged response times, particularly for complex topics like 'Llama Indexing and GPU Utilization' and 'Agents and Tool Interactions'. In contrast, established fields like Mobile Development and Security enjoy quicker resolutions and stronger community engagement. These findings highlight the need for improved community support and targeted resources to assist LLM developers in overcoming the evolving challenges of this rapidly growing field. This study provides insights into areas of difficulty, paving the way for future research and tool development to better support the LLM developer community.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.020 | 0.110 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.004 | 0.002 |
| Scholarly communication | 0.004 | 0.009 |
| Open science | 0.001 | 0.005 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".