This Week in Programming: Microsoft Makes Low-Code No-Code with OpenAI GPT-3 Model
Beware the sensational headlines forewarning the end of programming jobs as we know them, for they are far from true. Nonetheless, for all of us who might dabble, who might consider ourselves to be “the programming equivalent of a home cook,” our lives are about to get even easier with the news coming out of this week’s Microsoft Build developer conference, where the company announced that it would be harnessing the GPT-3 model to “help users build apps without needing to know how to write computer code or formulas.”
The news builds on another announcement from earlier this year, when Microsoft unveiled its low-code Power Fx language at Microsoft Ignite. Power Fx, dubbed “the low-code programming language for everyone” by Microsoft, is essentially an open source formula language for low-code that’s based on Microsoft Excel used in Microsoft’s Power Apps. With the introduction of this new feature, low-code quickly becomes nearly no-code.
This is huge for non-technical people:
Write in English, get database queries in return. @Microsoft already integrated this GPT-3 powered feature in some of its products.
I love how ML is becoming a tool for democratization! pic.twitter.com/pyTbwfzlAm
— Javier Fuentes Alonso (@JavierFnts) May 27, 2021
Now, as you can see above, we’re not talking about AI writing an entire application for you, but rather translating a natural language query into a database query. The example Microsoft offers is turning “find products where the name starts with ‘kids'” into “Filter(‘BC Orders’ Left(‘Product Name’,4)=”Kids”).” Rather than relying on documentation to find the correct commands and syntax or, let’s face it, Google or StackOverflow to find a piece of pre-written code of someone else’s to CTRL-C and CTRL-V, these developer-hopefuls can now type their query in English (that’s the language of choice for launch) and hopefully get a query that suits their needs.
Microsoft’s phrase of choice, by the way, is “citizen developers,” and they explain that said “citizen developers” will still need a basic understanding of logic to use the new feature.
“The features don’t replace the need for a person to understand the code they are implementing but are designed to assist people who are learning the Power Fx programming language and help them choose the right formulas to get the result they need. That can dramatically expand access to more advanced app building and more rapidly train people to use low code tools,” they write.
As for how the whole thing works, GPT-3 is a “massive natural language model that runs exclusively on Azure” originally developed by OpenAI for which Microsoft has an exclusive license to integrate directly into its products. GPT-3 scans billions of pages of publicly available text to train its models and, in case you haven’t seen, can offer up some pretty convincing text, wherein an AI tries to convince you, dear reader, that it is not out to end humanity. It might not, however, have the rosiest picture of Muslims after all that scanning. All that aside, it is likely perfectly suitable for turning your natural language business language into database queries using all appropriate parentheses, commas, and semicolons.
I’m looking forward to hearing about all the new command injection vulnerabilities from natural language interpreters. The next decade is going to be a trip.
— will weld data pipelines for food (@eineBanaleFrage) May 26, 2021
And lest you remain fearful that the robots are here to take your jobs, Microsoft ends its announcement with a quote from Charles Lamanna, corporate vice president for Microsoft’s low code application platform: “In all cases, there is a human in the loop,” Lamanna said. “This isn’t at all about replacing developers, it’s about finding the next 100 million developers in the world.”
This Week in Programming
- And Then There’s The Rest of Build 2021: Of course, there were a string of other Microsoft-related developer announcements at Build this week which we won’t spend time detailing individually, but rather direct you along to the correct locations. For example, Microsoft offers a Windows Developer’s Guide to Microsoft Build 2021 that offers a recap of all the recent innovations the company has made (many of which we have detailed here, such as the Windows Subsystem for Linux support for graphics processing unit (GPU) and for for Linux GUI applications in WSL, among many others). There’s also the recap of all things Azure, which includes new AI functionality, a slew of Java features, the introduction of the Azure Synapse Link for Microsoft Dataverse, and more. And really, if you want to find out about everything launched or updated this week, just head on over to Microsoft’s Build 2021 Book of News to find everything collated in one convenient location, summaries, links and all.
Writing some actual Java for the first time in a long time. pic.twitter.com/Zel4QzhyZ6
— James Ward (@_JamesWard) May 24, 2021
- How Green Can Software Be? It’s likely not your first thought as a developer, but perhaps it’s well time that it makes the top ten: how sustainable is your software? The most common example given these days is that of Bitcoin and its exceedingly compute hungry process that is said to have an impact comparable to thousands of hours of YouTube binge watching. Well, Accenture, GitHub, Microsoft and ThoughtWorks have joined up with the Linux Foundation to launch the Green Software Foundation, which it says has three goals: establish green software industry standards, accelerate innovation, and drive awareness and grow advocacy. Already, they say that data centers account for 1% of global electricity demand and this is expected to rise to 3-8% within the next decade.
- Roadmaps For The Year Ahead: Looking at the year ahead, there were two roadmaps released this week we thought noteworthy. First, Jetbrains offered up nine highlights from the Kotlin Roadmap, which looks at plans for 2021 and beyond. The Kotlin public roadmap has all the details, but the blog post distills it down to just these nine, including the new compiler, the bet on the WebAssembly, the new and experimental Kotlin/Native garbage collector, support for Apple Silicon, improvements to IDE performance and stability, new core libraries features, and more. Next, there’s the newly released Visual Studio 2022 Roadmap, which follows up on the preview the company teased last month, which both delighted and disturbed its users with the announcement that it would be going 64-bit. Visual Studio 2022 will have three key themes, they write: personal and team productivity, modern development, and constant innovation. For all the details, head on over to the Visual Studio 2022 announcement blog. High level highlights include the move to 64-bit, increased accessibility, improved diagnostics and debugging, and more.
Seen in SF pic.twitter.com/boUAtzUAZh
— Sam Lambert (@isamlambert) May 28, 2021
- Amazon Extends Lambdas, Offers ECS Anywhere: Finally, Amazon has come out with a few bits of news this week worth mentioning. First, Amazon ECS Anywhere is now generally available, bringing Amazon’s Elastic Container Service to locations outside of an AWS Region – such as on premise. Amazon ECS Anywhere lets users “run and manage container-based applications on premises, including virtual machines (VMs), bare metal servers, and other customer-managed infrastructure” with all the bells and whistles you get with Amazon ECS itself. Next, AWS Lambda Extensions are also now generally available, which brings monitoring, observability, security, and governance tools to the serverless architecture. The news comes with a number of new partners – Imperva, Instana, Sentry, Site24x7, and the AWS Distro for OpenTelemetry – and enables asynchronous responses, which “enables extensions to perform activities like sending telemetry to a preferred destination after the function’s response has been returned.” While there are many partners with extensions available, you can also build your own extensions using the Lambda Extensions API.
DevOps: Let’s run K8s so we can converge services on machines and save $$$
Also DevOps: lemme put a 100MB sidecar service mesh container in every pod instead of 300kb libgnutls
— Avishai Ish-Shalom (@nukemberg) May 27, 2021