Using larger AI models to teach smaller models allows for the creation of specialized models that fit specific operating domains or cost requirements.

The future landscape of AI may involve a combination of large models and smaller, specialized models that work in parallel, leveraging different areas of expertise.

Hassabis argues that while smaller, specialized models can be effective, the development of large, general models is essential for achieving significant advancements in AI capabilities. He also mentions that larger models are essential for training the smaller, more efficient models that we use extensively.

To build efficient models, we need to distill knowledge from larger teacher models, which allows us to tackle significant problems in science and medicine today without waiting for artificial general intelligence (AGI).

In medicine, hallucinations from AI can be life-threatening, so we ensure our models are trained exclusively on peer-reviewed medical knowledge to maintain accuracy.

OpenAI's recent paper shows that models can effectively grade medical answers based on specific criteria, indicating their potential in other domains.

There are numerous smaller-scale AI projects, such as those in agriculture and food production, that require a different approach than the traditional model of hiring large teams of engineers for development.

We built OpenEvidence by assembling a team of PhD-level scientists from top institutions, focusing on creating specialized AI models trained on peer-reviewed medical literature.

OpenEvidence's approach to training AI models is distinct; it does not connect to the public internet, ensuring that the information provided is based solely on vetted medical research. Additionally, OpenEvidence has formed a strategic partnership with the New England Journal of Medicine, which is unique as they typically do not allow AI companies to train on their research.