As we gather more data on user preferences, we can create highly personalized AI models that cater to individual needs, making interactions more effective.

In the future, we will see a proliferation of different AI models, with companies using the largest models for performance optimization.

To scale Arena, we aim to capture a diverse user base across various industries, enhancing our ability to evaluate AI models effectively.

As AI evolves, we need to shift from static benchmarks to real-time evaluation systems that can adapt to user feedback.

The AI tools we build need to continuously deliver additional value as the baseline for foundation models improves.

As AI becomes more integrated into various applications, we need to optimize for specific internal use cases, which may lead us to use other models if they better serve those needs.

We have a lot of data on how people learn on Duolingo, and while we use it to build models, we can probably do more with it to enhance AI progress.

We work with both large and small model providers, aiming to assist anyone who wants to collaborate with us on AI evaluation.

We work with both large and small model providers, aiming to assist anyone who wants to collaborate with us on AI evaluation.