What if your developers could code at the speed of thought? How would that change the way you deliver technology?
With the AI tools for developers that are now entering the market, enterprises are starting to realize big dreams like efficient digital product delivery and fast time to market.
They’re gaining on a future where instead of competitors disrupting them because time-to-value and reprioritization are too slow, they’re bringing products and services to market within months instead of years to quickly out-innovate competitors.
Join Dr. Mik Kersten, CTO at Planview, and Scott Densmore, VP of Engineering at GitHub, in dreaming about this realistic vision for the future in this Mik + One podcast episode.
Who is Scott Densmore?
Scott leads the team behind GitHub Copilot, the AI developer tool for pair programming.
Scott has had an amazing career building products that enable developers to do their best work, from his childhood building tools in basic to IDEs to data centers, powering developer workloads. Developer happiness and productivity are his passion. GitHub Copilot is now in the hands of millions of developers and has been reshaping how software is built.
Best clip
How GitHub’s Copilot revolutionized programming. In this video, Scott explains how GitHub’s Copilot enables developers to stay in the flow state that increases their productivity—and happiness.
Key takeaways
GitHub’s Copilot is refounding programming with AI
In case Scott’s work hasn’t gone before him, he’s been on a journey to revolutionize developer productivity for decades. He and the team at GitHub are using AI to provide a tool—that’s Copilot—which makes pair programming (to work in tandem with an expert developer) available for every developer.
They’ve spent months developing Copilot and training it on numerous coding languages. Copilot starts with code completion but goes far beyond that to save developers time. In fact, their tests showed that programmers using Copilot were 50% faster in creating a project with the same results.
Beyond Copilot, their goal is to “refound on AI”—meaning, they’re infusing the entire platform with AI and doing it in a way that makes developers more productive. But they’re not putting AI everywhere. They’re looking for where AI will be useful for customers and putting it there.
“At GitHub, we’ve always been about developers, and we’ve always been about productivity, and we’ve always been about developer happiness because that flow state is where developers find themselves more happy. So Copilot is that next evolution of that. And we’ve taken advantage of this new Generative AI to actually make that happen.“
Happier developers lead to more customer value
Scott points out that developers tend to jump jobs when they’re not happy with their team or organizations. This isn’t ideal even for enterprise organizations because onboarding developers isn’t cheap.
The solution lies in increasing developer happiness, and a big factor in developer happiness is enabling more productivity. It turns out that (most) developers don’t just want to hand code all day. Per Scott: “We pay developers to think, not to type.” Implementing solutions like GitHub’s Copilot that removes toil from their day creates productivity gains.
Ultimately, though, the end goal of AI for development productivity isn’t just to make developers happy. It’s to enable developers to bring more ideas, leading to more value for customers. Scott’s perspective: “Don’t think of AI as replacing people. It’s to increase your productivity to get more value.” Any obstacle you can take out of the development process improves your end-to-end value stream and gets more value to customers.
“As an organization, you should be thinking about that entire flow: how do I reduce the toil, how do I increase the value that I can give to my customers internally and externally, and how can I use these tools to make that happen?”
Even with AI, code will never go away
Even though GitHub Copilot uses AI to make coding faster, Mik and Scott agree that code and coders will always be necessary. Engineers have continuously tried different ways to get rid of code, and AI makes it seem like that’s in view. But while AI can communicate with natural language prompts, it’s simply another way to abstract the code. It still requires developers who understand how it works.
“Code is that means to an interface with that computer to solve that problem. But if we can make that easier to do and make it more productive for people to do, there will still be code behind that until we get rid of the zeros and ones.”
Explore more about AI and developer productivity
Mik and Scott continue the conversation about developer productivity and how this fundamentally changes organizational value delivery. Here’s what else you’ll get in this episode:
- In the value streams at enterprise organizations, just 8% of the time is spent in development. Hear what Mik and Scott think can address this problem.
- Why Mik and Scott are both encouraging their kids to learn code.
- Scott’s perspective on what this productivity revolution will do for organizations.
Want more?
Check out the Mik + One archives to listen to episodes with these and other experts. Be sure to subscribe to learn when new episodes are released.
- Jean-Michael Lemieux, former CTO and SVP of Engineering at Shopify
- Robin Yeman, Senior Fellow and Agile DevSecOps Enterprise Coach at Lockheed Martin
- Paul Littlefair, CTO, and BMK Lakshminarayanan, Value Stream Architect at Bank of New Zealand
