Jason Kingdon: Do Software Robots form a candidate for Moore’s Law?

Image: jeffedoe/Flickr

Gordon E. Moore of Moore’s Law made his iconic 1965 prediction that computer power will continue to double approximately every two years to persuade engineers that the more stuff they crammed on to a chip, the cheaper general electronics could become. It wasn’t meant to be precise or accurate. Coming true was as much of a surprise to him as it is to us.

The prediction has proved so useful and seductive that it is now being applied to science itself. The idea puts forth the question: Can the rate of scientific discovery be predicted? Intriguingly, many commentators have suggested that anecdotally not only is the answer “Yes,” but also that the general rate of discovery is increasing.

Tim Berners-Lee has said that the pace of innovation is escalating and that something needs to “happen” to break what is a clear and accelerating trend. Andrew McAfee and Erik Brynjolfsson, two MIT economics professors, similarly suggest we are entering a new historic phase where exponential increases in innovation will take hold. From another quarter, writer Milan Kundera has proposed that history itself is speeding up and that people will no longer live in one “period” but two, or possibly more.

If true, it would mean that on some vast scale Moore’s Laws are being created everywhere. The pace could even posit a form of “Rip Van Winkle” syndrome in which the inflicted feel amazed and bamboozled on stumbling over an innovation they had failed to notice.

A 20-year snooze from 1990 in IT alone gives us mobile communication, the internet, tablets, self-driving cars, chess-champion machines, social networks, online shops, YouTube, Skype, Google, and alternative living worlds. All major deals.

In business, such a syndrome is clearly lethal. It’s another form of business leader paranoia – “Is something being created now that could end my business tomorrow?” A fear Andy Grove, as a technologist, took very seriously.

One way to spot a Moore’s Law effect is if the innovation feeds-back on itself to produce accelerated return. A so-called singularity. Better chips design better chips, for example. A question I have been asked recently is: Do Software Robots form a candidate for Moore’s Law?

Software Robots are software that mimics a human. They drive other computers and software products the way that humans do. As such, they do not need specialist integration or IT skills. They are like “virtual” colleagues that cost a fraction of a current global wage and are aimed at replacing humans by doing routine, rote computer-based clerical tasks.

Do they feed-back on themselves? Well, this is intriguing. The current experts in large scale clerical work are the global business process outsourcers (including all forms of off-shoring). Richard Jones, one of Accenture’s original outsourcing pioneers, says he would typically target a third costs reduction for process execution: each year of a three year term.

This actually implies that, even with humans, the initial process improvement goal is an exponential impact. So using Software Robots must mean some acceleration. However, reducing cost exponentially like this will of course eventually converge to a limit, so for continual cost reduction, major innovation will also be needed.

One observation that may point to the way Software Robots offer a form of “universal” transport layer is between different software applications. They provide the means by which systems can be cross-patched without the need for systems integration.

Traditional system integration (making one system “talk” to another) has been responsible for more project glitches and postponements than any other form of technical hurdle.

Software Robots do away with this issue (albeit with some loss of speed) because they use the same user-interface as people. No dedicated APIs or specialist connectors are required, the robot simply bridges the existing applications in the way a user would (a virtual swivel). This capability opens the possibility of a large scale “network effect” across the complete IT estate — which is very new.

It is also enormously liberating — it erodes the single supplier syndrome that has been the hand-brake of most IT-based business innovation over the last 30 years. Even more than this, it offers the possibility of “integrating” whole supply chain IT systems. Imagine even asking an IT department for a quote on how much that would cost! Something that would have been impossible 10 years ago is reduced to a configuration exercise.

The impact is almost impossible to predict. Releasing a 30 year-old business hand-break will have profound consequences which odds-on will deliver a Moore’s Law style return precisely because it allows work to “flow” in new and previously unimaginable ways.

Another possibility is robots being used to configure other robots. This is much more obviously in the spirit of a Moore’s Law acceleration, and again there is no reason why this could/does not happen (user manuals, regulatory updates, workflow configurations, and user instructions already trigger cascading updates through syndicated robots).

Fast forward and this creates its own Van Winkle effect — awaking to find offices in the cloud “staffed” by robots would at the very least be a jolt. A point that has not been lost on the BPOs is the question as to whether Software Robots are the next outsource destination. Here, too, the smart money says “Yes” — the economics are too compelling for it to be otherwise. The sector’s current leaders may recall Andy Groves’ famous advice.

Jason Kingdon is chairman of the robotic automation company Blue Prism

Source: Wired


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s