Microsoft Build 2018: Top Five Takeaways
It’s a busy week with the Digital Transformation Forum kicking off in Chicago tomorrow, but I managed to slip in a quick trip to Seattle to this year’s Microsoft Build developer conference. It’s hard to believe that it’s been four years since Satya Nadella took over as CEO, but time flies. We detected early on in Satya’s tenure that there was a renewed focus on developers, and today it seems more the case than ever. Here are my top five takeaways from this year’s Build:
- Windows as an operating system and Office as a suite evolve toward a Microsoft Platform. For as long as I’ve been writing code, we’ve had multiple types of Microsoft developers. There are Windows devs, .NET devs, Office devs, and Azure devs, just to name a few. Sometime these dev segments overlap, but often they don’t. At this year’s Build, Microsoft is putting down a pretty large marker toward converging the first three developer segments listed above into a single “Microsoft 365” platform. The shift is marketed subtly, but it represents a strategic move to Windows as one part of the platform, as opposed to the firmament of it.
- A steady maturation of edge IoT capabilities. At the end of 2017, we predicted a spike in deployment of edge IoT solutions driven by manufacturing and security needs. Satya Nadella kicked off the Build keynote by noting that intelligence is dissolving into the fabric of the cloud and edge devices, but for IoT developers, the devil of distributed intelligence is in the implementation details. Microsoft’s answer: open sourcing the Azure IoT Edge runtime and its inference engine, along with Azure Sphere: a secure microcontroller for edge devices. Behind IoT Edge is a consistent edge-core programming model that take full advantage of containers and serverless functions. Demos of a DJI drone and Qualcomm camera using Microsoft’s edge runtime show that Microsoft is well on its way to giving developers a compelling IoT development platform.
- Expanded support for modern development of cloud-native apps. Over the past few years, Microsoft has steadily modernized its development tools to work with modern, cloud-native apps and services. Examples include Visual Studio App Center for continuous delivery of mobile clients, and DevOps projects as a first-class citizen in Azure. DevOps support includes deep support for container-based microservices via Azure Kubernetes Services (AKS). Visual Studio also gets a slew of updates, but two of my favorite are Live Share, which supports pairing and peer review of code, and IntelliCode, which adds AI into the developer’s workflow, providing inline suggestions of what algorithms or functions might be appropriate to call next.
- AI and ML that is ethical and consumable. At Build, the dystopian view of artificial intelligence took a back seat to a focus on near-term opportunities for augmented intelligence, coupled with the ethical responsibilities developers have when producing and consuming models. Examples of AI used for accessibility include an alternative view of our future — helping blind, deaf, and autistic users better communicate in powerful ways. Real-time transcription of in-person meetings and identification of tasks and to-dos are powerful examples of how AI can augment human capability instead of replacing it. Other examples of developer-consumable AI and ML include updates to the Bot framework for Azure Conversational AI, integration of cognitive services in Azure Search, and prebuilt models for speech text and computer vision.
- Interop, interchange, and integration. At scale, a cloud-native world is multi-cloud, multi-platform, multi-service, and multi-agent. Building cloud-native customer experiences will require developers to spend a lot of time integrating code, lest customers end up getting locked into a single platform vendor’s walled garden. Weaving through Build keynotes were many examples of integration in action, from support of ML model interchange formats like ONNX to integration of Cortana into Amazon’s Alexa to Azure IoT Edge support for Linux devices to event-driven integration at massive scale using Azure Event Grid and Azure Logic Apps.
There’s a lot at this year’s Build for developers to unpack. My overall takeaway: Today’s Microsoft is an increasingly confident organization that is accelerating adoption and support for cloud-native technologies. As a result, it increasingly feels like the company is charting its own course versus playing fast follower to other cloud and digital platform vendors. Given Microsoft’s long track record of taking complex development tasks and simplifying them, developers would be well advised to consider Microsoft’s new platform services and development tools for their cloud-native efforts.