Today’s computer industry is represented by a wide choice of software packages available on a number of platforms. The past 10 years have given major computer groups the possibility to improve the distribution of applications developed for their operating systems. Along with this improvement, the popularity of smart mobile phones has led to surge in the number of applications available. Now, the trend is towards converging mobile systems with systems installed on personal computers. A certain number of general and specialized APIs are available to developers to create programs on these devices—mobile or not—now present on the market. Today, several million applications are available to users, on every kind of platform. Adding to the converging of mobile and non-mobile systems is the trend toward interconnecting applications with services in the name of facilitating the user experience.
Several APIs for application development
At this point for developers, things are becoming more complicated. Along with the APIs now used to develop applications on available devices are other APIs used to connect to services on the web in order to increase the application’s value-add. To add to this, new companies are offering “smart APIs” to create a general aggregation around a simple application, offering vocal interaction with users for example. This multiplication of APIs makes development increasingly complicated, and the dynamic of the ecosystem in which the applications operate increasingly obscure. What’s more, a subject that is taking on increasing importance these days has not even been addressed: personal data. Users are rightfully wondering how many different APIs and companies will be using their personal data. Many companies now record users’ actions on their applications in order to improve these users’ experience. This does not bode well. A system of untrammelled global surveillance is to be avoided.
Towards a new ecosystem to respond to the users’ need
Like any computer model or architecture that has become too complicated, the services and applications ecosystem offered to lambda users seems due for an update and, above all, a new architecture. Users’ needs are centred mainly on knowledge, information and actions—getting news on a region of the world, gathering material on a writer, ordering movie tickets, making a specific home automation command or processing an Excel file by interpreting them in a certain way with a view to making a presentation. The list goes on. To fulfill these needs, users are now adapting to multiple applications, nearly none of which is interrelated. These applications, such as the ones specified above, are becoming increasingly complicated to develop. It’s clearly an opportunity to enter a new era, one of artificial cognition: the fusion of data, algorithms and language.
A robot to answer any questions
This change is considerably altering the application ecosystem currently in place. Instead of a store furnishing applications to users, these users can now find a robot in the Cloud, a kind of autonomous assistant capable of understanding natural language to act on behalf of the person using it. Users will no longer have to look for applications to see which ones are compatible with their needs. Instead, they’ll be asking the robot if it’s able to perform such and such an action. The developer’s role will be just as important as before, however the emphasis of his work will be different. They will no longer be required to develop an application to affect users. They will be able to directly integrate functionalities into the robot and develop new behaviours for the robot. These new behaviours will integrate into the robot’s experience, increasing its knowledge and scope. There will be an API to obey, naturally, but only one. The interconnection of functionalities and behaviours will happen automatically.
Cognitive system as a service
This type of cognitive service will raise the level of abstraction when there’s a problem to solve requiring massive integration of data. The goal is get programmers and users concentrating on knowledge rather than technical problems. In this way, problem solving assistance will have found a new way of processing: a complete integrative system.