IoT (Internet of Things) has been the buzzword of the past year, and connected devices are now popping up everywhere. That leads us to the question of whether developments in IoT, AI and cloud are bringing us closer to the digital singularity in 2019. It definitely seems like a step in that direction. I’ve purchased a smart refrigerator that sends me push notifications (including a photo!) when the eggs are not placed in the right spot. You might be wondering: how useful is this really? It is fun though.
The cool thing is that my refrigerator is getting smarter because I assess the photos that I receive and provide feedback. A 5G network would enhance these forms of edge computing (another buzzword). It applies to all the ‘things’ that cannot be connected to a normal network.
An important consideration is and will continue to be how these devices are secured. It seems as if the primary focus is on developing smart products while security is an added bonus. How smart is that really? This has led to incidents in recent years where smart thermostats, security cameras, and even toys have become the target of cybercriminals.
I should also add that I am not afraid my smart refrigerator will be hacked, and for outsiders to see that my eggs are not in the right place. But that’s also because I have added a little extra protection of my own.
Cloud: the second wave
In recent years, we have witnessed the first wave of companies adopting the cloud. In general, they were young, data-driven companies with a lot of technical knowledge. They developed their own software, knew their way around APIs, and hardly needed any support from their cloud provider.
Now, we are seeing a second group of companies adopt the cloud. These are the multinationals that run on SAP and have their own data centers where they run their accounting systems and all kinds of legacy applications. Some of these companies migrated to AWS – and they were shocked because that turned out to require considerable care and financial support. Many organizations have chosen to take a more gradual route. As a result, there’s still room for immense growth in the cloud market in the coming years.
For the first wave of cloud adopters, the following period will be all about optimization. They will focus on improvement by embracing hybrid and using cloud services in a smarter way. Many organizations will deploy multiple clouds and gain better control of their costs. The second wave of companies are faced with the following choice: head straight for the cloud or find a data center provider that also supports cloud solutions in order to support the new additions to their IT landscape.
Automation: taken over by the cloud
In 2018, we saw a huge push towards serverless solutions and using cloud technology to replace work previously carried out by developers and system administrators. This is part of a widespread push for more automation and an effort to replace work with technology systems. Virtualization has eliminated the need for system administrators to run to their servers, while containers ensure that applications are less dependent on the base level infrastructure.
Function-as-a-Service (FaaS) is the next step in removing complexity from programming applications. One element of FaaS involves trying to replace development with automatically generated code. The developer can then focus on the functionality within a FaaS platform, while the ‘plumbing’ is completely taken care of. Considering the scarcity of highly-skilled developers, the use of such cloud-based services is becoming increasingly attractive.
AI: speech recognition services are learning
Another obvious trend that we will be hearing a lot about this year is AI. If we are to believe the market hype, AI will be the ultimate solution for everything. However, most of the solutions appearing on the market are entirely unrelated to artificial intelligence. AI simulates thought patterns, with the key characteristic that AI can improve its thought patterns based on assumptions from the past. That’s the beauty of intelligence: it makes itself smarter.
The vast majority of the solutions that are presented on the market as AI do not actually ‘learn’ at all. They work on the basis of pure pattern recognition. For example, a chatbot will answer based on it recognizing some words typed by the user. However, if you enter something like, ‘I didn’t mean that. My tracking code shows that my package has been delivered, but I didn’t receive it.’ … then the bot does not suddenly give its answer.
In 2019, we will see a shift from relatively simple solutions based on pattern recognition, to real AI models and solutions.
Siri is a good example of this. If you say something to Siri today which it doesn’t understand, the data is sent back to Apple using a built-in feedback loop, ensuring the pattern will be updated accordingly. Because of that feature, voice recognition services are getting smarter, since they are trained continuously with new data.
It took my smart refrigerator a couple of months to realize that the eggs were in fact oranges. They were still in the wrong place, though.