In the recent issue of the New Scientist, 70 very smart people predict the biggest scientific breakthroughs of the next 50 years. The topics range from the evolution of consciousness and the universal theory of everything to implantable, organo-electrical brain-machine interfaces and the discovery of extraterrestrial life. In the computing field, here are a few of the predictions:
Google Director of Research Peter Norvig speculates on the future of search, and envisions digitial intermediaries taking care of business. Sounds like the Apple Knowledge Navigator concept come to life:
We are in the middle of an expansion of information access, with the internet providing democratic access to billions of pages of text. Most of this is mediated by search engines.
The only other comparable expansion started in 1456, with the introduction of the printing press. Fifty years and 15 million books later, the theologian Sebastian Brant wrote "There is nothing nowadays that our children... fail to know."
Today, 12 years into the era of search engines, we still have not made good on Brant's boast. Search engines deliver relevance but knowledge requires human work. In 50 years the scene will be transformed. Instead of typing a few words into a search engine, people will discuss their needs with a digital intermediary, which will offer suggestions and refinements. The result will not be a list of links, but an annotated report (or a simple conversation) that synthesises the important points, with references to the original literature. People won't think of "search" as a separate category - it will all be part of living.
Ray Kurzweil is quite a bit more aggressive in his prediction, which comes straight from his book, The Singularity Is Near: When Humans Transcend Biology. Obviously, Kurzweil's views have stirred up controversy.
By 2020, computational power sufficient to simulate the human brain - about 10 million billion calculations per second - will be available for $1000. The software will take a decade longer, but we are making exponential gains in reverse engineering regions of the brain. As a result, by the late 2020s, the tool kit we use in artificial intelligence will include all the processes involved in human intelligence.
I have consistently predicted that by 2029 we will be able to create machines that pass the Turing test. The result will be a formidable combination, uniting the subtlety and suppleness of human intelligence with the ways in which machines are already superior - for example, in their ability to download knowledge at electronic speeds.
The combined intelligence of all human brains is relatively fixed at 1026 calculations per second. By 2056, our nonbiological intelligence will be a trillion times greater than this in terms of hardware, and will also be vastly superior in software due to many generations of redesign. But this will not be an alien invasion of intelligent machines: rather, we will merge with the tools we are creating.
In the next 50 years, computer science needs to achieve a new unification between the inside of the computer and the outside. The inside is still governed by the mid-20th century approach – that is, every program must have defined inputs and outputs in order to function.
The outside, however, encounters the real world and must analyse data statistically. Robots must assess a terrain in order navigate it. Language translation programs must make guesses in order to function. Because the interface to the outside world involved approximation, it is also capable of adjusting itself to improve the quality of approximation. But the inside of a computer must adhere to protocols to function at all, and therefore cannot evolve automatically.
A unified computer science will be able to produce computers with internal structures that connect via approximation instead of precise protocols. It is reasonable to guess that this unification will be bio-mimetic, and symbiotic to improved understanding of the brain.
See the full New Scientist special report with 50 predictions for the future