seth_lloyd's picture
Professor of Quantum Mechanical Engineering, MIT; Author, Programming the Universe

I think less. My goal is to transfer my brain's functions, bit by bit, to the Cloud.

When I do think, I am lazier. There's no point in making the strenuous trek over to the library to find the source when you can get an expurgated electronic version on Google books right away. And why go look up the exact theorem when you can find an approximate version on Wikipedia?

OK, you can get burned. Math being what it is, an approximate theorem is typically an untrue theorem. Over the years, I have found most statements in purely scientific reference articles on Wikipedia to be 99.44% correct. It's that last .56% that gets you. I just wasted three months and almost published an incorrect result because one clause in the Wikipedia statement of a theorem was, in fact, wrong. It's a lucky thing the referee caught my error. In the meanwhile, however, I had used one of the great Internet innovations, the scientific preprint archive, to post the incorrect result on the Internet for everyone to see.

For hundreds of millions of years, Sex was the most efficient method for propagating information of dubious provenance: the origins of all those snippets of junk DNA are lost in the sands of reproductive history. Move aside, Sex: the world-wide Web has usurped your role. A single illegal download can propagate more parasitic bits of information than a host of mating Tse Tse flies. Indeed, as I looked further afield, I found that it was not just Wikipedia that was in error: essentially every digital statement of the clause in the theorem of interest was also incorrect. For better or worse, it appears that the only sure way to find the correct statement of a theorem is to trek to the library and to find some book written by some dead mathematician, maybe even the same one who proved the theorem in the first place.

In fact, the key to correctness probably does not even lie in the fact that the book was written by that mathematician, so much as that the book was scrupulously edited by the editor of the series who invited the mathematician to write the book. Prose, poetry, and theorems posted on the Internet are no less insightful and brilliant than their paper predecessors: they are simply less edited. Moreover, just when we need them most, the meticulously trained editors of our newspapers, journals, and publishing houses are being laid off in droves.

Life, too, has gone through periods of editorial collapse. During the Cambrian explosion, living systems discovered the evolutionary advantage of complex, multicellular forms. Like the digital organisms of today's Internet, the new Cambrian lifeforms rewrote the rules of habitat after habitat, evolving rapidly in the process. Finally, however, they filled their environment to its carrying capacity: at that point, just being cool, complex, and multicellular was no longer enough to insure survival. The sharp red pencil of natural selection came out and slashed away the gratuitous sequences of DNA.

For the moment, however, the ability of the Internet to propagate information promiscuously is largely a blessing. The preprint archives where scientific work (like my wrong paper) are posted for all to read are great levelers: a second- or third-world scientist with a modem can access the unedited state of the art in a scientific field as it is produced, rather than months or years later. They, in turn, can produce and post their own unedited preprints, and so on. As long as computer memories keep doubling in capacity every year or two, those stacks of unedited information will keep doubling and doubling, too, swamping the useful and correct in a sea of extraneous bits. Eventually, the laws of physics themselves will stop this exponential explosion of memory space, and we will be forced, once more, to edit. What will happen then?

Don't ask me. By then, the full brain transfer to the Cloud should be complete. I hope not to be thinking at all.