Re:90% of patients respond to new blood cancer treatment in ...
by Terr_ on Monday June 26, 2023 @11:05PM
(#36488418)
Potential issues:
1. To be definitely safe, you can't just sample one healthy cell, you have to sample lots and from the many different types and subtypes everywhere in your body. Imagine training a swarm of robots to attack anybody wearing a mask and wielding a knife... and then they enter a hospital surgical ward.
2. Sometimes the problem with cancers is that they are behaviorally wrong in ways that aren't easily expressed at the cell-boundary. (Your body has MHC-1 proteins that try to offer a debug-window into what the cell has been doing, but that's its own complicated topic.)
1. To be definitely safe, you can't just sample one healthy cell, you have to sample lots and from the many different types and subtypes everywhere in your body. Imagine training a swarm of robots to attack anybody wearing a mask and wielding a knife... and then they enter a hospital surgical ward.
2. Sometimes the problem with cancers is that they are behaviorally wrong in ways that aren't easily expressed at the cell-boundary. (Your body has MHC-1 proteins that try to offer a debug-window into what the cell has been doing, but that's its own complicated topic.)