Intel has set up an interesting program at its Intel Labs, the company’s hub for experimental projects with long-term prospects. The Machine Programming Research (MPR) project will attempt to automate software development for complex platforms while reducing coding errors. The primary reasoning behind the project appears to be the increasing shortage of trained or expert programmers who can write complex software code reliably and consistently.
Intel Labs’ MPR project will use multiple forms of machine learning and other automatic methods to create a platform capable of creating its own software mentioned Justin Gottschlich of Intel Labs who has been asked to lead a team of machine programming researchers. The ultimate goal of MPR appears to be developing a completely automated platform that’s capable of writing software and code with minimal human intervention and errors. However, the project appears to be intended for people who are unable to code. The MPR should eventually help people with no programming skills, but ample creativity, to build their desired software or platform without ever writing a single line of code themselves.
Intel Labs Gets Machine Programming Research Makes Lofty Promises About Realistic Problems In Software Development:
Software is deeply integrated within every aspect of modern civilization. However, developing and maintaining software is a time-consuming and error-prone process, notes Justin. He claims “Machine Programming”, wherein the power to create and maintain software is in everyone’s hand, could significantly simplify the complex process. The main problem within the Software Development segment is the availability of dedicated, experienced and reliable senior developers.
Simply put, there’s a severe mismatch between the growing requirement of the industry, and the ability of the real world offering skilled professionals. Justin observes in the near future it will become increasingly difficult, perhaps impossible, “to find developers who can correctly, efficiently, and securely program across all of that hardware.”
Intel is researching machine programming: "MPR uses forms of machine learning and other automatic methods to create software capable of creating its own software. It will enable everyone to develop their own software without writing a single line of code." https://t.co/PTGb9i2EnT
— witeken (@witeken) October 28, 2019
Machine programming uses multiple automatic programming techniques that range from precise (e.g., formal program synthesis) to probabilistic (e.g., differentiable programming) methods. It also uses and learns from everything Intel and other hardware manufacturers have developed in hardware and software to date.
The two most needed aspects of software development are reliability and consistency. As Windows 10 OS updates have clearly and routinely indicated, accuracy and performance improvements almost always come with bugs and weird behavioral patterns. MPR aims to investigate and flush out common bugs, errors and other problems that regularly find their way into the software. The next step would obviously be to fix them automatically, observed Justin.
Intel’s vision of programming in the future: Machines will do the programming. We humans only provide our intentions. pic.twitter.com/g3ofonxOkq
— Primeur Magazine (@primeurmagazine) June 25, 2018
Along with reducing common bugs in software, Intel’s MPR program will also work to improve and eventually reduce the amount of code that goes into software. By eliminating lines of code, the software would become leaner, cleaner and more efficient. Moreover, with less number of code within the software, it’s accuracy would improve as well.
It might appear that Intel is out to eliminate Software Development jobs. However, Justin categorically noted that the Intel Lab’s MPR program is not to reduce the workforce. On the contrary, MPR is poised to create multiple new jobs and work opportunities. Moreover, the core agenda of MPR is essentially to free up the professionals to take on more complex components of software development. This is simply because MPR would eventually handle the menial, run of the mill aspects of programming. With the entry-level blocks of software being taken care of, programmers could focus on better integrating multiple platforms like machine learning and formal techniques, heterogeneous hardware, and many programming languages into the software platforms.