At Assembled, engineering velocity is paramount, requiring robust testing to maintain quality without slowing down. Large Language Models (LLMs) like OpenAI’s o1-preview and Anthropic’s Claude 3.5 Sonnet have revolutionized test generation, reducing tasks that took hours to just 5–10 minutes. Engineers have saved hundreds of hours reallocating time to feature development. By using LLMs to enhance testing practices, teams can maximize productivity. Using precise prompts, engineers can generate detailed unit tests efficiently. The iterative refinement, smart model choice, and custom prompts maximize the benefits of LLMs in test generation, leading to a cleaner, safer codebase with increased development speed.
https://www.assembled.com/blog/how-we-saved-hundreds-of-engineering-hours-by-writing-tests-with-llms