Anyone else spending more time fighting the model than doing the actual work?
Users are reporting that large language models like Claude and ChatGPT struggle to maintain context and adherence to instructions over multiple iterations. This often leads to a degradation in output quality, requiring users to either repeatedly re-enter prompts or manually edit the generated text. The issue appears to be a common frustration for those relying on these models for professional tasks such as content creation. AI
IMPACT Highlights practical challenges in using current LLMs for sustained, complex tasks, suggesting a need for improved long-term context management.