You need big models to train smaller models effectively, and while we have efficient smaller models, they often rely on insights from larger models.

We often underutilize models, expecting them to solve tasks quickly without giving them enough time to iterate and improve.

When developing a new large language model product, it's essential to iterate on prompts to find what works best for your specific use case.

It's crucial to analyze quantitative data, like user logs and retention curves, to understand how your product is actually being used, rather than relying solely on user feedback.

Using large language models as prototyping tools can be effective; they allow for rapid development and testing of ideas before creating more specialized models tailored to specific tasks.

Evaluating the performance of applications built with large language models is challenging because the use cases are often subjective, making traditional accuracy metrics insufficient.

It's beneficial to observe users interacting with the app to identify areas for improvement and enhance usability.

It's important to observe how people use software in real time, as this can provide valuable insights that metrics alone cannot capture. Relying solely on metrics to understand user behavior is like observing raw data from a radio telescope instead of experiencing the night sky directly.

AI models need to develop a deeper understanding of users beyond single interactions, fostering empathy and contextual awareness to enhance long-term engagement.