The future will consist of many smaller, specialized models for different use cases, rather than just one generic model.

The future will likely see more researchers embedded in every product team, as fine-tuning models for specific use cases becomes a core workflow for building products.

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.

While specialized models can be effective for specific tasks, the overall trend is towards general models that incorporate learnings from specialized iterations.

The ability to distill models from larger ones will be valuable, as companies will likely have AIs tailored for every function they need.

The current landscape shows that large model labs are focused on building powerful foundation models that require less fine-tuning, which may lead to a decline in the relevance of fine-tuning in the near future.

The first generation of AI applications in drug design often focused on local models, but our approach is to build a more comprehensive system that can generalize across various tasks, as demonstrated by AlphaFold 3, which allows us to create models that generalize across chemistry and target space, enabling application to any protein or small molecule without needing fine-tuning.

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

There's enormous potential in specializing general techniques for specific applications, such as protein structure prediction, and we are actively exploring this in our AI for science work.