Virtualization Technology News and Information
VMblog Expert Interview: Mathew Lodge Talks Diffblue Cover and Artificial Intelligence for Software Development


With almost every organization in the world accelerating their investments in digital transformation (a lesson reinforced the hard way during this terrible global COVID 19 pandemic), velocity in writing software is quickly becoming a critical competitive market advantage in any industry. I recently caught up with Mathew Lodge, CEO of Diffblue, who thinks artificial intelligence can pick up much of the grunt work required in creating code so companies can write and ship code faster with fewer defects. His company today is announcing the general availability of their commercial Diffblue Cover as well as Diffblue Cover: Community Edition, a free version created for developers using IntelliJ, the most popular IDE for enterprise Java.

VMblog:  Your offering seems to be one of the first of its kind in the industry, basically using machines to write code.  You use artificial intelligence as the engine of your solution.  What's the story behind your company?

Mathew Lodge:  More software needs to be written in the next ten years than there are people to write it. The company was founded in 2016 after we saw a huge multi-billion dollar market opportunity to apply AI to help solve this problem. Why should high-priced software engineers spend up to a third of their time doing drudge work like writing tests if there was a way to automate that? Unit tests are the foundation for high performance DevOps pipelines because they find problems at the time the code is being written. So we are starting with unit tests and focusing first on Java given its immense popularity for enterprise software, though the underlying technology will also be compatible with with any programming language.

Diffblue Cover automates writing unit tests and we can do that 10X to 100X faster than a human. The unit tests Cover writes are also easy for developers to understand, and Cover automatically maintains the tests as the code evolves - even on applications with tens of millions of lines of code. Our free Community Edition makes it easy for Java developers using IntelliJ (the most popular Java IDE) to discover how much automating unit test writing can free up their time for more valuable and interesting (to them) problems.

VMblog:  How does artificial intelligence work in your solution?

Lodge:  The underlying AI approach is called reinforcement learning. It's the same machine learning strategy that powered AlphaGo, Alphabet (Google) subsidiary DeepMind's software program that beat the world champion player of Go. Our co-founders come out of academia, Professor Daniel Kroening from Oxford University and fellow researcher Dr Peter Schrammel. Daniel is a world-renowned expert in artificial intelligence, program analysis and program synthesis. He and Peter have been researching this space for over two decades.

VMblog:  Why is it important now that companies start automating much of how they write their software?

Lodge:  The short answer is competition. With the rise of the cloud, SaaS startups are disrupting legacy industries in every vertical. Even cloud giants like Microsoft Azure run enormous amounts of Java code. Enterprises need to be agile and move much faster in their business. They can't let software challenges slow them down, especially with these new upstart challengers. And they have piles of legacy code not properly unit tested, a drag on the race to digital innovation.

Let's look at financial services. There are hundreds of millions of lines of code written in Java that run the biggest banks in the world. That software is doing everything vital to the bank, like real-time pricing for derivatives and securities and much more. I'm talking now about production code, not just new greenfield applications. Much of that code is not well-covered by unit tests today, which makes it hard to ship quickly and frequently. If you are going to keep up with your competitors, then you need to be able to iterate and change the software constantly, shipping new versions much faster. Big players like Google and Apple are moving into financial services, and you also have to compete with fintech upstarts like Monzo and Transferwise. Every industry vertical is being disrupted by cloud-native challengers and their legacy competitors who moved to digital transformation ahead of them.

VMblog:  Do you have customers already?

Lodge:  We were fortunate to work early with customers such as Amazon Web Services (AWS) and Goldman Sachs. We have other big-name customers in other verticals today. We learned a lot working closely with these companies. Cover is production ready for any Java developer today. In fact, Goldman Sachs was so impressed (here is a link to our case study with them) that their venture arm joined with Oxford Innovation Sciences (the University of Oxford's technology venture arm) to fund our Series A in 2017, at that time the largest Series A round for an AI startup in the U.K.


About Mathew Lodge

Mathew has over twenty-five years of experience in the software industry in developer, product and marketing roles. Before joining Diffblue, his titles included SVP at Anaconda and Vice President of Cloud Services at VMware. In each role, his focus is on building and marketing products that customers love.

Published Tuesday, September 08, 2020 7:45 AM by David Marshall
Filed under: , ,
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<September 2020>