What do you think of a computer which can learn and reprogram itself? IBM flagged this with its Watson system, but things are moving fast. The game show playing Watson is now being replaced by serious computing capabilities.
Watson, in fact, was the forerunner of a big move beyond “normal” computing. Dr. John E. Kelly III, IBM Senior Vice President and Director of IBM Research has a lot to say on his blog about the new wave in computing, and it’s all new:
Cognitive systems are capable of learning from their interactions with data and humans—essentially continuously reprogramming themselves. Traditional computers are designed to calculate rapidly. Cognitive systems are built to analyze information and draw insights from it. Traditional computers are organized around microprocessors. With cognitive systems, it’s about the data and drawing insights from it through analytics.
This is the big shift away from abacus-type computers to truly efficient computing. The fact is that while the existing systems are OK for what they are, they have some serious limitations. Code becomes obsolete fast. It can do what it can do, and can't really change, despite endless patching and upgrades. Coding is hard work, and the more complex the code, the more issues that code creates. Top code writers can be considered to some extent totally wasted on number-grinding roles. Cognitive computers represent a massive asset to coding as a science. This will be real generational coding, not “fire and forget” coding.
(A slight digression, but to illustrate an important point- What happens if you let a cognitive computer learn endlessly? You’d get, at the absolute least, an entire track of code covering a real evolutionary process in coding. That alone makes cognitive computing a very respectable idea.)
Computers which can reprogram themselves according to need are also likely to be efficient research and development tools. To progress, a process involving blundering into an issue, finding the solutions and writing the code for them is less than sparklingly efficient. A computer which can not only reprogram itself but define the entire process and its solutions on an ongoing basis is about as far ahead of current computers as a bird is from a rock as a flying machine.
This has been talked about for a very long time, in fact since the Golden Age of Science Fiction, but it’s never really been a functional possibility before, let alone a working reality. Your home computer or equivalent, with cognitive capacity, could find its problems, solve them and tell you what it did. You could interface with it and ask it to deal with software problems or just get it to make your games play better.
Cognitive computer security is also likely to be a very different ball game. Cognitive computers, in theory, could predict malware functions before they happen, simulate them and fix them before they’re even operational. Few hackers would be able to match them.
Do cognitive computers pose a threat to humans?
All technology, from fire and tool use onwards, can be and historically has been abused. The other side of this rather oblique and often “niche” argument is that without them, there would be no civilization. All technologies create risks. A computer which can analyze your behavior and predict what you’ll do next is a potential threat in the same sense as an invader of privacy. Humans naturally predict each other’s actions, rightly or wrongly.
Could cognitive computers run your life for you, whether you like it or not?
Cognitive computers, as assets, could be used to streamline your life and make it more efficient and rewarding. The learning process Kelly refers to is perhaps the most directly valuable of cognitive computing’s potentials. Can it learn how to balance your budget for you? Probably. Can it learn how to make your work more productive? Quite likely- If it can access and analyze the relevant information and apply it to your needs. Yes, they could run your life, directly or indirectly.
(It’d be interesting to see how a cognitive computer approaches workplace issues like productivity and job design. My bet is that it would be nothing like the current farcically unscientific efforts.)
There’s an interesting point to be made here- One of the Holy Grails of computing has always been exactly this sort of computer. The mighty Omnivac of Isaac Asimov’s stories was the omniscient ruler of the world, based on vacuum tube technology, but a machine which could solve the biggest problems. Douglas Adams’ Hitchhiker’s Guide to the Galaxy included the famous Deep Thought, which even found a way of making the Ultimate Answer to Life, the Universe and Everything useful.
Kelly makes it clear that cognitive computers aren’t “thinking computers”. This is number crunching with an ongoing purpose, and that’s a very different ball game. The difference is that this is also going to be as big a leap as personal computers, probably bigger, and faster.
There’s also a certain irony which won’t be lost on either technology buffs or the public- Cognitive computing eliminates the “software molestation” which has produced so many ridiculous techno-turkeys and absurdly overweight operating systems and software packages. Cognition involves facts only. Facts don’t have agendas, and nor does good mathematical analysis.
The cognitive computers will call it like they see it. That will be a very nice change.
****Readers please note- Do read Kelly’s blog post. It’s a very first glimpse of a new era.
This opinion article was written by an independent writer. The opinions and views expressed herein are those of the author and are not necessarily intended to reflect those of DigitalJournal.com