Carl Zimmer has an excellent analysis of of the Flores Hobbit brain. At the current pace I might as well dump Carl's RSS on to GNXP, but, in any case, the highlights:

• The probability that it is microcephalic goes way down.
  
• The probability that it is a H. sapiens pygmy also seems to have decreased.
  
• The probability that it is some form of H. erectus seems to be about the same.
  
• The probability that it is "something else" seems to go way up (mostly at the expense of microcephaly and human pygmy hypothesis).