16.11.18

Why 2019 will be the inflexion point

stage and back of crowd shot at Samsung's CES press conference

Let’s face it, the ‘smart’ paradigm doesn’t work anymore…

Peter Aylett, CT at dubai installer Archimedia, black and white headshotThese days, it’s like saying something is ‘electric’. Well DOH, yes, move along now, nothing to see here. Our industry tries to build on this with control systems that use conditional logic, which we then try and claim is ‘smart’. Really? Industrial automation has been using PLCs (Programmable Logic Controllers) since the 1970s, which is functionally little different from the kinds of stuff we do these days such as ‘IF Sun = 30 minutes before setting, THEN turn on a lighting scene’. Is this smart? I don’t think so…

Today, all the marketeers are getting on the bandwagon with the AI (Artificial Intelligence) paradigm without really knowing what that means. AI comes in many different flavours with the vast majority of current applications such as voice recognition being very narrow in scope as opposed to true broad AI, which aims to replicate and eventually exceed human intelligence across all its nuanced facets.

Whilst narrow AI can be, and is mostly fantastic at mastering one specific task, it’s still not anywhere near the level we need to move to the next paradigm of technology in buildings, which is proactive assistance. To get there, AI boffins need to build the smart machines, but it’s the machine learning eggheads that will make those machines truly intelligent.

Machine learning, which Stanford University defines as “the science of getting computers to act without being explicitly programmed,” needs three things; context, input, and properly clever programming. Context comes from a huge database of facts (think human – our entire life’s experiences). Input comes from sensors of every type (think human – we have five senses giving us the input to be intelligent and learn). The clever programming comes from the clever programmers now leveraging almost unimaginable computing processing power born out of gaming and now being nurtured largely in autonomous vehicles.

So, who’s going to crack the new broader AI paradigm of proactive assistance in buildings? Will it be companies dedicated to our channel? Unlikely, but they will help with one critical element. Will it be the likes of Samsung or LG? Watch this space as my prediction is that 2019 will be the year when the big CE companies stop treating ‘smart home’ as a proof of concept hobby and start playing for real.

“…2019 will be the year when the big CE companies stop treating ‘smart home’ as a proof of concept hobby and start playing for real.”

At Samsung’s CES press conference in January this year, the hour was taken up mostly with them showing their AI driven ecosystem of devices all around a building. Was it contrived, set up and massively rehearsed? Of course, but it was no less impressive for it. Is there an ‘our channel’ company who could even get close to the level of proactive assistance shown by Samsung? Not even close!

So, here’s the thing – whilst the specialist manufacturers in our industry channel can help with the one critical element of input from sensors, with the other clever bits of context and programming they haven’t got a chance in the face of the double-digit billions of dollars being thrown at the problem by the big companies. We’re already seeing experiences being delivered with wireless mass market devices all seamlessly connected within one of the many available ecosystems eclipse what is possible with even the most sophisticated programming on the best industry control systems. 2019 will see an acceleration of this meaning that unless we adapt and move on from our old big rack full of control processors and matrix switches past, we as an industry will die a quick death whilst hearing the laughter of the big CE companies echo through the halls of CES.

There will always be a place in niche ultra-custom and ultra-secure projects for sandboxed control systems that rely on no cloud communications or processing. For everything else, 2019 will change everything and will force all of us to accept the new reality that to achieve broad AI proactive assistance in buildings, the future is in mass market consumer products.

Peter Aylett is president and CTO of Dubai-based integrator Archimedia

 

www.archimedia-me.com