3 Comments
User's avatar
Claus Wilke's avatar

Maybe similar: AI people think they can just train their models on massive biological datasets without having to understand the biology or biochemistry of how the data was generated, and biologists think the AI has magically figured out all the biology and biochemistry and can now answer any question an experimentalist may have.

Expand full comment
Cristhian Ucedo's avatar

The obvious solution is polymaths. We need a lot more of them!

Expand full comment
Nathaniel Joseph's avatar

interesting post … things like this lead me to believe the optimal amount of “epistemic trespassing” is not zero

Expand full comment