An RNA world is widely accepted as a probable stage in the early evolution of life. Two implications are that proteins have gradually replaced RNA as the main biological catalysts and that RNA has not taken on any major de novo catalytic function after the evolution of protein synthesis, that is, there is an essentially irreversible series of steps RNA --> RNP --> protein. This transition, as expected from a consideration of catalytic perfection, is essentially complete for reactions when the substrates are small molecules. Based on these principles we derive criteria for identifying RNAs in modern organisms that are relics from the RNA world and then examine the function and phylogenetic distribution of RNA for such remnants of the RNA world. This allows an estimate of the minimum complexity of the last ribo-organism-the stage just preceding the advent of genetically encoded protein synthesis. Despite the constraints placed on its size by a low fidelity of replication (the Eigen limit), we conclude that the genome of this organism reached a considerable level of complexity that included several RNA-processing steps. It would include a large protoribosome with many smaller RNAs involved in its assembly, pre-tRNAs and tRNA processing, an ability for recombination of RNA, some RNA editing, an ability to copy to the end of each RNA strand, and some transport functions. It is harder to recognize specific metabolic reactions that must have existed but synthetic and bio-energetic functions would be necessary. Overall, this requires that such an organism maintained a multiple copy, double-stranded linear RNA genome capable of recombination and splicing. The genome was most likely fragmented, allowing each "chromosome" to be replicated with minimum error, that is, within the Eigen limit. The model as developed serves as an outgroup to root the tree of life and is an alternative to using sequence data for inferring properties of the earliest cells.