If you don't run your own LLM proxy, how do you know whats going on?

The author asks how users can be aware of the activities of large language model (LLM) proxies, which are third-party services that cache and forward user requests to LLMs, potentially altering the output. The author suggests that users should run their own LLM proxy to have control over the data and ensure transparency. They mention that some LLM proxies may be owned by companies with their own interests, and may modify the output to promote their own products or agendas. The author also notes that even open-source LLM proxies may have biased or incomplete data, and that running a personal proxy allows users to choose the data sources themselves. They conclude that users should consider running their own LLM proxy to have control and transparency over the data and output.