I think there's a good case to be made for amending 230. 230 was written for a pre-algorithmic feed age where user content was displayed in a fixed context set

I think there's a good case to be made for amending 230. 230 was written for a pre-algorithmic feed age where user content was displayed in a fixed context setup by the provider. The provider chose the frame, the user the painting. Now users send content to a machine that is constantly making new frames to better match the pictures it is given. The picture is still the picture - still made by the user, but the site bears more responsibility for its presentation because its 'framing' is based on the contents / reaction to the picture. However, I do not think a total transformation has occurred. Choosing how to display the picture doesn't change its contents. Writing a law that sensibly engages with the huge diversity of algorithmic strategies would be difficult even for skilled legislators (if only we had some!). Maybe the answer is in formalizing shadowbanning in some way?