The Fat Alberts Man, that is. Just came back from hearing Bruce Alberts speak. He talked in general terms about his early career, his work on Molecular Biology of the Cell, his time as president of the National Academy of Sciences, and most recently about his work as editor of Science magazine. He had some interesting things to say, although nothing really ground-breaking. He generated a lot of good will by talking about his early graduate work and about how it was a complete failure, and how he literally failed his thesis defense and was required to go back and do another six months of research. One of the things I found most interesting, though, was his repeated insistence that young scientists need to come up with new, novel ways of doing things, and not just repeating the type of work that others have done before them, and about how he personally wanted to use the journal Science to encourage that sort of thinking. But this seemed to me to be highly ironic. Journals like Science and Nature, have a system of pre-peer-review editorial review, where they reject a large fraction of papers for not being interesting enough, or sexy enough, or relevant enough, irrespective of the quality of the science. This is actually the complete antithesis of what he seems to be striving for, and I think it's well accepted (at least among people that I know) that major journals play a pretty active role in discouraging experimentation on innovative systems with innovative techniques, because the journal editors just won't understand why it's interesting or important if it strays too far from the known corpus.
Compare this with the talk I heard at the Science Commons salon recently, by Peter Binfield, the managing editor of PLoS ONE. PLoS ONE operates in exactly the opposite fashion: they will publish anything, as long as the science is sound. They then rely on the community to rate the work based on relevance, importance, or whatever else they deem interesting, sort of like Digg or Reddit. The question, of course, is whether the Intarwebs-at-large is a better judge of what's interesting, useful, and innovative than the editors of Science. I'm not really sure. I'm not even sure that "interesting" is a useful metric. But, "interesting" is what interests people, so it will always have some relevance.