Update README.md

README.md

@@ -48,13 +48,14 @@ And they are generally (but not always) generated in the order above, for which
 
 For models less than 11B size, I experimentally generate f16 versions at the moment (in the static repository).
 
-For models less than 19B size, imatrix IQ4_NL quants will be generated, mostly for the benefit of arm.
+For models less than 19B size, imatrix IQ4_NL quants will be generated, mostly for the benefit of arm,
+where it can give a speed benefit.
 
 The (static) IQ3 quants are no longer generated, as they consistently seem to result in *much* lower quality
-quants than even static Q2_K, so it would be s disservice to offer them.
+quants than even static Q2_K, so it would be s disservice to offer them. *Update*: That might no longer be true, and they might come back.
 
-I specifically do not do Q2_K_S, because I generally think it is not worth it, and IQ4_NL, because it requires
-a lot of computing and is generally completely superseded by IQ4_XS.
+I specifically do not do Q2_K_S, because I generally think it is not worth it (IQ2_M usually being smaller and better, albeit slower),
+and IQ4_NL, because it requires a lot of computing and is generally completely superseded by IQ4_XS.
 
 Q8_0 imatrix quants do not exist - some quanters claim otherwise, but Q8_0 ggufs do not contain any tensor
 type that uses the imatrix data, although technically it might be possible to do so.