LLMs + DNA = OMG?

Here’s a scientific blockquote for your Friday morning/afternoon/evening, courtesy of a blog post from Stanford Report:

Trained on a dataset that includes all known living species – and a few extinct ones – Evo 2 can predict the form and function of proteins in the DNA of all domains of life and run experiments in a fraction of the time it would take a traditional lab.

And here’s another:

Imagine being able to speed up evolution – hypothetically – to learn which genes might have a harmful or beneficial effect on human health.

Apparently, scientists have created this generative AI tool called Evo 2 to help in biological sciences and they think it could reshape it too. It’s open source which sounds nice(r) but then there’s the possibility of someone taking elements of it and making a for-profit version (even if the licence demands referencing of the original which some companies don’t do and still secure clients and funding).

Now “trained on a dataset that includes all known living species” doesn’t specifically mean the dataset has yours or my DNA (I hope???) but the seeing the words “DNA”, “run experiments”, and, “imagine being able to speed up evolution – hypothetically […]” makes me cautious given how awful the US government is and who is in charge of health over there.

I’m not trying to fearmonger or give absolutes here but stuff like this doesn’t fill me with hope and excitement when those who fund this kind of work have harmful agendas. The paper trail is never a good one when it comes to AI or genetic science (lest we forget Francis Crick who advocated a form of positive eugenics in which wealthy parents would be encouraged to have more children).