Some users are experimenting with running large language models (LLMs) locally, rather than relying on cloud-based AI, with mixed results. Local LLMs appear to have strengths in areas such as coding, automation, and brainstorming, but may struggle with more complex tasks. It is unclear whether local LLMs can fully replace cloud-based AI models, with some users reporting limitations and others finding them suitable for specific uses.