The technology industry has no choice but to embrace innovation and take risks. Thus, some innovations start out looking crazy or eccentric, but end up being brilliant. On the other hand, there are those that may not yield any return at all.

In this light, here are six technology ideas that walk a fine line between the amazing and, well, what some might consider stupid. The developers of these innovations could turn out to be crazy - or they could become insanely exceptional. The technology could end up being a black hole for venture money or a clever play for business value emerging along the margins. It all depends on your perspective.

Quantum computers

Of all the existing technologies, nothing is more hyped than quantum computers - and nothing is more terrifying. The work is done by a mix of physicists and computer scientists tinkering with strange devices at super-cold temperatures. If it requires liquid nitrogen and lab coats, well, it has to be innovation, right?

The potential is enormous, at least in theory. The machines can work through zillions of combinations in an instant, delivering exactly the right answer for a mathematical version of Tetris. It would take millions of years of cloud computing time to find the same combination.

Cynics, however, point out that 99% of the work we need to do can be accomplished by standard databases with good indices. There is very little real need to look for strange combinations and, if there is, we can often find perfectly acceptable approximations in a reasonable period of time.

Cynics, however, are still looking at life through old-fashioned glasses. We haven't even started asking the questions that quantum computing can answer. Once the machines are readily available, we can begin to think about asking the new questions. That is one of the reasons why IBM is offering quantum computing toolkits - and certification for those who want to explore the outer limits of what the machines can do.

• Potential early users: Domains where the answer lies in searching for an exponentially growing combination of hundreds of different options.

• Likelihood of happening in the next five years: Low. Google and IBM are at war with press releases. Your team will spend many millions just to reach the press release stage.

Heating with computing

Every decision made by a CPU sends a few electrons down the ground wire and whatever energy they carry is turned into heat. Traditionally, these joules have been treated as waste, and finding a way to get rid of this heat has been a headache for circuit designers and computer case builders.

The other approach is simply to use less electricity, a strategy that could work should questions arise about green energy. (Are windmills killing birds? Are dams killing fish?) Instead of asking algorithm designers to find the most amazing algorithms, just ask them to find the simplest functions that get close enough. Then ask them to optimize that approximation to put the least load on the most basic computers. In other words, stop dreaming of mixing a million-layer algorithm trained by a dataset with billions of examples and start building solutions that use less electricity.

The real secret strength behind this movement is the alignment between innovation and environmentalists. Simpler calculations cost less money - and use less electricity, which means less stress on the environment.

• Potential early users: Casual AI applications that may not support expensive algorithms.

• Potential for success in five years: High. Saving money is an easy-to-understand incentive.

Build your own cloud clusters

Yes, they are tiny computers that cost less than $50. Yes, some 4th graders are setting them up for science fairs. But just because they are cheap toys doesn't mean they can't be very useful for real work. That is why some are building Raspberry Pi clusters with racks filled with tiny Linux nodes using quad-core ARM chips that sip electricity, not drain it.

There are many reasons to avoid this idea. Big, fat machines can be much more efficient. They can offer dozens of cores running dozens of threads and sharing large blocks of RAM and disk packs. When loads get heavy, they can provide power during the work.

But working with smaller, separate machines offers redundancy precisely because they are separate. You might think your instance is separate from other virtual machines, but they often share the same CPU, and there could be dozens or even hundreds of them. Separate machines with separate circuit boards offer security and redundancy.

The biggest win, though, might be the price. These clusters can be far, far cheaper than some of the instances in the major clouds. Sure, some machines in the cloud cost only $5 a month, but after a year, the Raspberry Pi might start to look cheaper.

Clusters like these allow massively parallel algorithms to run freely. Many of the most intriguing problems require churning through massive data collections, and often, the tasks do not need to be performed in order. These machines allow programmers not only to think about inherently parallel algorithms, but also to start building and implementing them.

The trend also follows the way some major clouds are adopting solutions that offer hybrid options to move data back on-premises. Some want to save money. Some want security. Some want assurance.

• Potential early users: Shops with big data sets that need parallel analysis.

• Potential for success in five years: High. Clusters are already being deployed.

Homomorphic encryption

The weak point in the world of encryption is the use of data. Keeping information locked down with a reasonably secure encryption algorithm is simple. Standard algorithms (AES, SHA, DH) have withstood continuous attacks from mathematicians and hackers for some years. The problem is that if you want to do something with the data, you need to unscramble it, and that leaves it in memory, where it is prey for anyone who can slip through garden-variety holes.

The idea with homomorphic encryption is to redesign computational algorithms so that they work with encrypted values. If the data is not decoded, it cannot leak. There is an abundance of active research that has produced algorithms with varying degrees of usefulness. Some basic algorithms can perform simple tasks, like querying records in a table. More complicated general arithmetic is trickier, and the algorithms are so complex that they can take years to perform simple addition and subtraction. If your calculation is simple, you might find that it is safer and simpler to work with encrypted data.

IBM, one of the leaders in the field, has been fostering exploration, offering toolkits for Linux, iOS, and MacOS developers who want to include the functionality in their applications.

• Potential early adopters: Medical researchers, financial institutions, data-rich industries that must protect privacy.

• Potential for success in five years: Varies. Some basic algorithms are commonly used to protect data. Elaborate calculations are still very slow.

Tricorders everywhere

Most of the technology in Star Trek remains a distant dream, but we have already grown accustomed to putting one of their so-called "communicators" in our pockets. If anything, the current generation of mobile phones is much sleeker than the flip phones Kirk and Spock would use.

Our society's next target could be the tricorder, the box that medical teams would wave in Star Trek to diagnose illnesses and scan our hidden innards. The good news is that the scriptwriters were never specific about what a tricorder does. We know a phaser could kill or be set to stun, but the tricorder was essentially a prop to occupy Dr. McCoy's hands before he said, "He's dead, Jim."

Some researchers are already spreading the word. One group is working on a “DNA tricorder” that will decode DNA sequences and fit in your pocket. Others have assembled a digital stethoscope, EKG sensor, lung sensor, and a blood sampler that pricks your finger. Qualcomm has awarded $10 million in prizes and defined a tricorder as a device that can capture five vital signs and diagnose 13 potential conditions.

But we can do more. Right now, CT and MRI scanners are large and expensive, requiring elaborate radiation emitters and super-cooled sensors. But point sources of radiation are everywhere in the form of cell phone towers. If a sensor for that radiation could be made with just a fraction of the sensitivity and resolution of digital cameras tuned to the visible spectrum, well, the computational power of GPUs should begin to make sense of our bodies' interiors. Signals from nearby cell towers or television stations could act as point sources attenuated by the various body tissues.

• Potential early users: Everyone from surgeons in operating rooms to first responders at the scene of an accident. Home users with chronic illnesses and hypochondriacs will be big fans.

• Likelihood of happening in the next five years: Low. It depends on what you think a tricorder can do. Some basics, like measuring blood oxygen, are simple and already on the market. However, detecting tumors buried in the pancreas will take longer.

Why not use it to heat buildings in the winter, then? Why not replace the world's boilers and heat pumps with miniature racks of servers pumping heat? The people living upstairs would be grateful. Computing jobs could migrate from North to South and back to the North with the seasons, much like Arctic terns that spend half the year in the Northern Hemisphere and half in the Southern.

In these cases, there would be some challenges. If a heat front arrived, say to New York in January, inhabitants would turn off the “heaters” and decrease the cycles available to AI researchers, data scientists, and everyone else who buys spot instances. It could also mean installing twice as many servers, or perhaps shipping servers was cheap enough.

Right now, cloud companies keep their servers in massive racks in central locations where electricity is cheap. If they move to homes, they can reuse the heat.

• Potential early users: Countries with cold climates.

• Chance of happening in the next five years: High. Pilot projects are already being tested around the world.

Green Artificial Intelligence

If the buzzwords “green” and “artificial intelligence” work alone, why not put the two together and double the fun? The reality is a bit simpler than the double hype might suggest. AI algorithms require computing power, and at some point, computing power is proportional to electrical power. The ratio continues to improve, but running AIs can be expensive. And electrical power produces tons of carbon dioxide.

There are two strategies to solve this. One is to buy power from renewable sources, a solution that works in some parts of the world with easy access to hydroelectricity, solar farms, or wind turbines.

Article by CIO

Did you like it? Follow CodeBit on our social networks (Facebook, Instagram and LinkedIn) and keep an eye on the CodeBlog. 

}