Mobile World Congress 2018: Machine Learning & IOT
‘Cause all the hype don’t feel the same next year’ – Drake
I’ve been thinking about how the most important skill over the past year has been the ability to separate hype from reality.
It truly feels like we have hit peak buzzword.
Now Peter Thiel founder of the of PayPal’s mafia has some sage advice when it comes to dealing with companies that throw around buzzwords, that is…. ‘run away as fast as you can’.
It’s not that Buzzwords are inherently bad for marketing purposes, it’s just that buzzwords can disguise and confuse what ideally a company would like to be done, compared with what can actually be done.
This year at mobile world congress involved a lot of separation between the two.
Here are some of the key technologies that I found with real application being rolled out now.
Machine learning is an application of the famous buzzword ‘AI’, which when coded correctly is a way for computers to intelligently access data and learn through observance, direct experience or instruction how to achieve the set objectives the code has been made for.
Over the past few years it has largely been a playground for those companies with a large number of data and data streams to to train the machine learning algorithms on to get results. Such as Facebook’s scarily accurate lookalike audiences or Waymo’s self driving car.
This year there were some incredible applications using limited data sets that show how the gains are trickling down from the Silicon Valley elite and now becoming accessible to a wider array of companies.
Take the Cambridge Consultants who managed to train a data set using just 8,000 images of master paintings (which could easily be scraped from Google) to allow anybody to also become a master painter.
Just check out what started as a horrible drawing below and was turned into a masterpiece using these algorithms….
Additionally image recognition can now be executed locally on a device accessible to most using the Raspberry Pi. Just check out what the Japanese company Idein were showcasing at mwc…
Once polished this brings a whole new set of opportunities to retail, defence and appliance industries that were out of reach before now.
It used to be you just had two or three devices (PC and Phone or Laptop) connected into the ether.
Today you can probably just count how many things you have connected in the ether.
Tomorrow maybe not so…
Although the terminology behind IOT has changed in recent years. The tremendous number of devices currently being connected into the right now cloud is mind-blowing.
And a lot of it is coming from this company ARM. If you own any major smart phone it’s highly likely that it’s being powered by one of their chip designs.
And now MBED from ARM is being used to connect everything else.
Connecting in IOT chips like MBED now allows you to track and manage whole host of new things that couldn’t be accessed before. Giving aircraft and medical devices owners alike preventative breakdown data, city councils massive efficiency gains, supply chain logistics companies access and on, using a logistical framework….
And it’s not just the ‘new’ tech that had people intrigued at mwc. Companies like Devolo who blew me away with applications of the lifeblood that supplies this data and connectivity using internet and wi-fi boosters. It shows just how much the core elements of tech are progressing, or must be to progress in order to enable the newer layers on top to operate as efficiently as possible.
Now both machine learning and IOT might not be able to save the world yet, the foundations are cleverly being built and it’s just as exciting, if not more so in this building out stage. And it’s in these real world applications where the action is.