Looking to harness the power of AI for developers? Azure is making it possible. Hosts Patrick Moorhead and Daniel Newman are joined by Microsoft Azure’s General Manager of Digital Apps and Innovation Mike Hulme, and Mandy Whaley, Partner Director for Azure Dev Tools on this episode of Six Five Media On The Road for conversation on harnessing AI to empower developers and accelerate innovation.
📻 Tune in for more on:
- The transformative role of AI in the software development lifecycle, powered by Microsoft’s leading developer platform
- Enhancements in GitHub Copilot, including new AI models and integrations, to streamline developer workflows
- GitHub Copilot for Azure: A gateway for developers to build, deploy, and manage Azure applications more efficiently
- Empowering new developers with AI app templates and in-editor guidance to simplify their entry into AI-driven development
- Strategic approaches for teams to adopt AI-assisted tools and maximize their development workflows
Learn more at Microsoft Azure.
Watch the video below at Six Five Media at Microsoft Ignite and be sure to subscribe to our YouTube channel, so you never miss an episode.
Or listen to the audio here:
Disclaimer: Six Five On The Road is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Transcript:
Patrick Moorhead: The Six Five is On the Road here in Chicago. We are at Microsoft Ignite 2024. It is a huge event all about AI at many different levels. I mean, you are looking at IT, you are looking at users, you are looking at developers, which makes the world go around.
Daniel Newman: Yeah, it’s really good to be here, Pat. Chicago’s hometown for me, spent four decades here. It’s November and so when you say where’s the welcoming moment of the city of Chicago? Great city, this is the last edge of time that we can be here before it gets really, really rough. But it’s a great place to be, and you are right? I mean, we have to expect this week to be all about what’s going on with AI, but putting it together, Pat, right? It’s not just anymore about the theme. It’s not about being cool, it’s not about being first, it’s about getting it right and helping enterprises drive outcomes.
Patrick Moorhead: It is, yeah. On many different levels. I mean, enterprises can consume software in so many different ways. I mean, they can consume it through SaaS products, but most applications for enterprises today are still cranked out by developers. And Microsoft has really been leaning into developers. I mean, it’s always been developers, developers, developers, but this new wave of AI developers really seeing some cool stuff, GitHub, VS Code, cool stuff like that. And I can’t imagine two people that I’d rather chat about this than Mike and Mandy, welcome to the show.
Mike Hulme: Thanks, great to be here.
Mandy Whaley: Yeah, happy to be here.
Daniel Newman: Yeah, it is really good to be here. And Pat, I think you framed that up pretty well. Now we do have to be realistic about where the world is at, and Microsoft has been really ambitious with Copilot. We know that GitHub, we know that the way apps are being developed, it’s changing meaningfully. And so maybe Mandy, Mike. Mike, I’ll start with you first, but the AI assist, AI’s impact on how we develop. We hear the entire continuum, Mike, we hear some people are like, we’ll never need developers again. I’m sure that won’t be the message tomorrow.
Mike Hulme: Absolutely.
Daniel Newman: And then you hear others that are kind of saying, look, it’s going to augment. And then other people are like, “Oh, it’ll never happen,” or “It’s never going to be good enough.” Talk a little bit about how that AI assisted evolution is taking place and how Microsoft sees this.
Mike Hulme: Yeah, absolutely. Well, thanks again for having me here. And so I would say almost more than any other area we’ve seen, the software development world is the one that’s been transformed through AI to date. Here we are. We’re at about a two-year anniversary for GitHub Copilot, and I would say that AI assisted development is really almost ubiquitous in the developer’s view. Some developers use some kind of tool in their tool chain, and I’m obviously very proud to say that GitHub Copilot is the one that is generally the one that people are using and is the most widely adopted. And so we look at things like developer productivity, can I code faster? And in general, developers are saying, yes, I can code about 55% faster. That’s a study we’ve been doing for a couple of years now, but really now we’re turning our lens towards code quality. What’s the acceptance rate, what’s the error rate? How are we able to deliver higher quality code with higher confidence, really bring that code into the system with fewer vulnerabilities and really deliver an application that’s really ready for production. And that, I think is the key when you look at this AI pair programming tool, again, keeping the developer in the loop, but really giving them the tools to move faster.
Maybe the last thing is really around developer experience. So moving fast is great, but what’s the quality of the experience for the developer? What’s their joy that they’re able to bring back into their world? I’ve been around long enough to know that this idea of needing to move as fast as you can and never having enough time to get through your backlog as a developer, that’s been a perennial problem. I think we really have an opportunity to move forward there and put our energy into the things that matter. We actually looked at the Copilot’s impact inside of our own software development team within Microsoft, and we found that actually because of the productivity that we created, we are generating code at a throughput rate that’s now the equivalent of adding 3000 developers to our developer pool. That means that you’re moving faster, we’re innovating faster as a company, we’re delivering some really cool stuff, but the developers are focusing on what matters and gets them really excited.
Daniel Newman: Sure. Absolutely. Mandy.
Mandy Whaley: Yeah. So I work on the Azure Developer tools team, and so we’re really focused on how this AI-assisted part can help with when you’re building applications on Azure. And so that’s really where we’ve been spending a lot of time, both in terms of how does Copilot fit into your VS Code experience? How are you using GitHub? How does that connect to your Azure resources? How are we bringing that all together so that you can build whatever kind of application you’re working on?
Patrick Moorhead: It’s funny, as analysts, we always need to watch out for the n equals one, but my son and his entire class is using GitHub Copilot to literally do everything that they do. And he showed me how it worked, and I did a mini study, but it is truly impressive. It’s just not something that a lot of fake AI or AI washing, it’s actually working and you guys, Microsoft being customer zero, the amount of code that you create, literally, I don’t know if there’s measurements on who does more code commits than any other company on the planet. I’m sure Microsoft is up there, but it is fascinating to see that one of the biggest, most accepted, and I would say even the most credible metrics for the benefits of AI is in the programming element. Anyways, hoping that spreads everywhere at this point. But I do have to ask, you talked about, we got it out there, initial focus was on efficiency and speed, and then the second one was code quality. What did you announce here at Ignite? What new features? What new capabilities did you announce here?
Mike Hulme: Yeah, well, you’re right. So the coding efficiency and speed code completion, that’s really just the beginning. That’s table stakes, and we love that that’s been accepted. To your point, it’s probably one of the most widely used uses for AI today. But again, it’s just the beginning. We really want to extend this same concept of bringing AI and AI assistance across the entire software lifecycle. So maybe on the front end, one of the things that we’ve done is we’ve made it really easy for developers to actually choose the model that they want behind GitHub Copilot. So whether that’s coming from OpenAI or if you want to go and work with another model of choice, you can actually now embed that into your Copilot experience and have the best model for your particular preference. So that’s number one. The second thing is also allowing you to do really complex code changes across multiple files.
So really looking at the full spectrum of things that you need to do in that one area and making it something that’s much more efficient. So that’s a great efficiency thing, but now when you look at what we’re doing around security, scanning for vulnerabilities, using AI assistance to remediate those vulnerabilities, but keeping you in the loop and giving you the option of accepting or modifying those changes as you go through it. We’re even helping on the app modernization side, doing things like Java upgrade assistance. Soon we’ll extend that to .NET and other languages, but really helping to keep applications up to date, keeping vulnerability and security up to date as well. All of that is about really enhancing the entire life cycle. Now that’s just in Copilot. There’s also some really cool things that we’re doing when you combine Copilot and Azure, and I know that’s really what a lot of people are excited about as well this week.
Patrick Moorhead: Yeah, for sure. Mandy, how does this intersect from your point of view here?
Mandy Whaley: Yeah, absolutely. So as part of working on Azure Developer tools, one of the things I love about working with Microsoft is we get to think about the whole developer journey, just the whole end-to-end experience from starting in your editor, where you’re writing your code through your DevOps tools like GitHub and Azure DevOps and all the way connecting to Azure and GitHub Copilot is there to help you that whole way. And so one of the things that we’ve been working on and that we’re showing this week is GitHub Copilot for Azure. So this is a Copilot chat extension that you can use inside of VS Code that helps you with all things Azure. It knows your Azure context, it knows about your subscription and your tenant, your resources. It also has all the latest up-to-date docs for Azure.
So if you’re learning about a new feature or something, maybe you heard about it, Ignite, you don’t have to go and search on your own. You can just ask GitHub Copilot for Azure directly inside of your IDE, get the information and start coding. And then it can also help you with things like troubleshooting applications, managing your resources, and building and deploying applications. So we’re really excited about this ability to help the developers stay in the flow in their editor, but have all the connections to Azure that they need through GitHub Copilot.
Patrick Moorhead: Yeah, I was going to say, it sounds like the easy button for this. That’s great.
Daniel Newman: Well, it’s really interesting though is because it doesn’t really matter if you’re a developer or any sort of knowledge worker context switching is problematic. We talk about where do we lose productivity? The chair swivels, a lot of people like to throw the chair swivels. Well, if you’re a developer and you’re moving from platform to platform, that’s substantial. It’s just like us as analysts, if we work on three different research papers at the same time, you’re bouncing from things. It takes a while to get readjusted to it. So it sounds pretty powerful. I mean, have you guys started to assess, Mandy, how much impact this has?
Mandy Whaley: GitHub Copilot for Azure is very new. It just came into preview right now.
Daniel Newman: Why don’t you know?
Mandy Whaley: Yeah, right. But one of the things that you were saying about the context and that context switching, we know that that has an improvement for developers when they’re able to stay focused on their code, but maybe they have a quick question about, I want to check this setting on my Azure resource. Now they can do that very quickly just in their chat panel, get right back to coding. And then the other thing that’s really powerful is that because you’re in your editor, you have your code open, you have your repo open, we’re able to combine the context of which application you’re working on, what language you’re working in, what resources it’s using with your Azure context, so we can give you the best help for what you’re trying to do based on all that context. So that I think is really powerful and will just continue to get more powerful.
Patrick Moorhead: In the preamble. Sometimes we have to state the obvious, which is how enterprises consume applications, and obviously building your own applications is a huge part about that. We talked about how AI could not only help you as a coder to be more efficient and you’re increasing the accuracy. And with Azure AI plugin another element of doing the easy button, but I’ll start with you Mike. How is this helping or how are you helping them add AI to their own applications beyond what we’ve talked about?
Mike Hulme: Yeah, it’s a great question. So I think there’s this really interesting tension that’s happening right now. There’s this really eagerness to move quickly. Everyone sees the opportunity with AI, they see the potential for just incredible impact to their business, but then they also have to be able to run these applications with confidence. Sometimes they’re working in totally new areas.
Patrick Moorhead: You mean the board of directors mandate does not know that that has to happen too?
Mike Hulme: Exactly.
Patrick Moorhead: Oh, my gosh.
Mike Hulme: The pressure from the top is often the biggest point of tension that we have because you know have to do something and you might even have a timeline that’s obviously totally unrealistic. But we also have to make sure that if we’re running this application and it’s generating revenue or it’s building and understanding data to help us innovate faster, or if we’re driving a customer experience that’s going to be out in the public and part of our overall brand, then we also have to make sure that it’s reliable, it’s secure, and it’s actually giving us the results that we need. So one of the things we’re bringing out this week is something called Azure AI Foundry and Azure AI Foundry really brings together the full suite of technologies across our AI portfolio, everything from our model catalog to our portal environment and all the features that people really love about Azure AI.
At the same time, it then exposes those services to developers. And one of the things it does is it helps every developer, if you’re working in GitHub or Visual Studio, you now have access to a full portfolio of models that you can start to experiment with. You can compare them for performance or for the type of results that you’re looking for out of those models, you can really easily pick the right model for the right application, and you can also start to work with other tools that allow you to build and deliver agents. So all of those are about bringing these AI capabilities, which may be new to many of the developers that we’re speaking to this week into the tools that they already know, they already love. There’s less context switching to our point earlier, but it’s also bringing it into environment that’s really familiar, and we think the learning curve is going to be much, much easier for people.
Patrick Moorhead: So this might be even applications that exist today and they want to give them superpowers. Is that right?
Mike Hulme: Yeah, that’s a great point actually. We talk a lot about the build new environment because it’s so interesting, but actually the opportunity to modernize the existing state of applications is really massive. There’s so many legacy applications or applications that are in market today that are part of core business processes, and one of the things we’re seeing is while the traditional use cases for AI like agent environments or customer service, customer support, those continue to be very popular. But we’re also seeing the rise of these business processes that need to be infused with intelligence to actually make them more modern, more successful, and obviously more valuable.
Patrick Moorhead: Okay. Mandy, how about you?
Mandy Whaley: Yeah, so one of the things we’ve been thinking about is how do we get developers hands-on starting to build their own AI-powered applications, and how can we get them there as quickly as possible because everyone’s very eager and ready to dive into this and start building things. So one of the things we’re announcing this week is the AI app Template Gallery. And so this is a gallery of templates that show you different AI application patterns and architectures. They’re written in all the different languages, Python, JavaScript, Java, .NET, and you can go and search by the task, the use case, the model that you’re using and each one of these templates then, it’s like a starting point that you can take and then get it up and running in Azure in minutes so that you can see it working, start to explore how it fits together and then start to modify and move from there.
All the templates have source code for the application. They have the infrastructure as code files and bias up for Terraform, so it’s repeatable deployment process. And you can work with them in VS Code, with GitHub Copilot for Azure. And a lot of them also include GitHub Action Pipelines, which is really an important part. So one thing we thought about is when you’re building an application, we all have all of our software development skills for building all the applications we’ve been building for a long time, but what’s this new AI part and how do we fit it into our CI/CD pipeline? How do we think about the prompt evaluations and things like that? So we’ve built that into some of these templates to give people a pattern that they can start working with the new SDKs for evaluations, setting up GitHub Pipelines, and have a working application that they can then start to modify really quickly. So we’re excited. We think that’s going to be a great place for developers to get started when they’re faced with, hey, we want to start building an app and moving towards pilot and production phases.
Daniel Newman: So Mandy is this, GitHub models is that-
Mandy Whaley: This is different than GitHub models. You could use it with GitHub models, but this is actually just a gallery of, think about it as sample apps-
Daniel Newman: Just to bring this to life. What’s a use case-
Mandy Whaley: They go from simple to more enterprise and advance. One simple one that’s great that we’re talking about this week in some of my sessions is getting started with GPT-4o and image processing. So that’s a super cool feature. A lot of people want to do that. We have an application that you can get from the template gallery, select where you want to deploy GPT-4o, deploy it, have it running up in about four minutes, start uploading images and chatting about the images. So that’s one that’s great. But they also go towards more complex like an enterprise RAG architecture where you’re ingesting your internal documents, you’re building an internal assistant that could help people with questions about internal business processes or things like that. And we also include things like managed identity and a lot of the security features that you’re going to want for those more enterprise type solutions.
Daniel Newman: Some of it like the low code platform though, in terms of out of the box, here’s the solutions, the power platform version for coders-
Mandy Whaley: In terms of giving you these starting points. But it is all code driven, so if your goal is to take it, and-
Daniel Newman: I used the bad word.
Mandy Whaley: Adapt it.
Daniel Newman: Talking about how someone that knows this stuff though can really piece things together much more quickly.
Mandy Whaley: Yes, exactly. Yeah. Some of them are what we call building blocks that you can put together, but they’re code building blocks that help you then take in and build that up. And it’s like sample apps with superpowers, to go back to the super power analogy.
Daniel Newman: So take me upstream now to GitHub models. So I feel like you’re heading in that direction.
Mandy Whaley: Yeah, GitHub models is great. So GitHub models gives you a way just natively in your GitHub experience to explore this really expansive catalog of all the different models that are out there. So there’s new models, so much innovation happening in the model space, and it’s really important to be able to experiment with those in a playground type setting to this, the first phase of building your intelligent app, it’s figuring out which model fits your use case, trying sample prompts, comparing different models. GitHub models has this great compare feature that lets you try out different use cases with the different models. And so that’s all in GitHub and super quick and easy to get hands on and start asking questions. And then once you’ve settled on a model and you’re ready to get into building your application, you can easily move to Azure AI Foundry from there and move over and then start taking your app towards more of a pilot or production phase.
Patrick Moorhead: I got it. That’s great. So it’s been a great conversation, but I always like at the end of these to what’s the one piece of advice you would give developers? Mandy, I’ll start with you. What advice would you give top number one, you don’t get two or three, to help developers pull AI tools into their workflows?
Mandy Whaley: Yeah, I would say that my top advice is you’re building on all the skills you already have. Those are important. You’re building applications, starting to integrate AI into your apps is mostly starting to make some API calls that start to bring that in. So be confident, know that you already have a great skill base for that, and all of the AI coding assistants can help you along the way. And my real advice is just start, get started. Build something.
Patrick Moorhead: I love it. Mike, how about you?
Mike Hulme: Yeah, so I think the first thing is whether it’s your son or someone who’s a career programmer, this idea of AI infusion into your applications, into your development tool chain, it’s really for everyone. No matter what your skills are, no matter what your knowledge base is. So what I would say is don’t look at it as just code completion and code generation. Look very broadly, how is this introducing you into new concepts and new tools? How are you now becoming more proficient in areas like models where maybe you haven’t had any exposure or to new platforms. Maybe you’re not an Azure programmer, but now you can be incredibly proficient in building reliable applications that are going to run in production in a way that you really intend them to. So these tools and these services are going very broadly, huge opportunity beyond maybe what we know in that narrow band of code completion. It’s really just the beginning. It’s getting more and more exciting every day.
Daniel Newman: It really is exciting.
Mike Hulme: Yeah.
Daniel Newman: This is great. And as we sort of pull this all together, we can see that continuum we started the conversation about AI assist, and the assist is basically people with limited beginner knowledge can start to think about how to add AI. People that are deeply experienced can move really, really quickly with the new tools and capabilities that are being rolled out here at Ignite. So Mandy, Mike, I want to thank you both so much for joining us here on The Six Five.
Mike Hulme: Thanks for having us.
Mandy Whaley: Yeah, thank you.
Daniel Newman: Let’s make it a great Ignite.
Mandy Whaley: Absolutely.
Daniel Newman: And thank you for tuning in and being part of The Six Five. We are here at Microsoft Ignite in Chicago, my hometown, hit subscribe. Join us for all of the other episodes here of The Six Five and all of our episodes for Patrick and myself. But for this episode, we got to say goodbye. See you all later.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
