In particular, we still lack robust, general answers to questions such as:
Will text-based AI eventually be able to reason over raw genomes, or is biological data too fundamentally different?
How much training data will biological models need before qualitatively new capabilities emerge, or they āgrokā?
What is the tradeoff between evading sequence-based screening and preserving biological function?
If a model was never trained on viral sequences, how well will it generalize to viruses from other domains of life?
In addition to these kind of general questions, it could also be valuable to just have a dedicated per-model evaluation to inform decisions BAIM creators do.
That is true. Iām a part of a team building a genome language model. Itās really unclear even to us which models are the best models! and a lot of these fundamental questions can also guide model design.
In that way, if we do it right, and thoughtfully, itās possible to both become a resource for model construction and improve their safeguards and security.
In addition to these kind of general questions, it could also be valuable to just have a dedicated per-model evaluation to inform decisions BAIM creators do.
That is true. Iām a part of a team building a genome language model. Itās really unclear even to us which models are the best models! and a lot of these fundamental questions can also guide model design.
In that way, if we do it right, and thoughtfully, itās possible to both become a resource for model construction and improve their safeguards and security.