Learn more about GPT-3 and how we use it at Crowdbotics
The GPT in GPT-3 stands for "Generative Pre-trained Transformer." This is the third iteration of the model, and the whole point of it is to produce text that sounds natural to humans. For example, GPT-3 can write news articles that sound like they were written by real people.
Crowdbotics is using GPT-3 to bridge the gap between natural language descriptions of a function of an application and the code needed to make it work. If we can make building software as simple as describing the way you're envisioning it in your head, we can truly open up the world of app-building to anyone and everyone. We're not quite there yet, but the engineering team here at Crowdbotics is excited to see what can be done with the updated capabilities of the GPT-3 model, and you'll see this reflected in our product.
To use GPT-3 in the Crowdbotics App Builder, you have two options:
1. Visual Editor
In the Visual Editor, you can use GPT-3 to generate actual screen layouts by describing the screen. Simply go to your Visual Editor, select the "Editor" tab at the top, and then you can access the experimental feature. Type into the layout description box and the app will generate a screen for you! Keep it simple. Once the code is generated, select “copy to clipboard." You now have the code needed for your new screen copied. Go back to the Visual Editor and make sure you're in the "Code Editor." This can be accesssed from the tab at the top of the tool. Delete everything in the screens/blank.js tab and then paste your copied code. You will see the new screen on the right.
2. Storyboard Tool
From the Storyboard tool, you can now use GPT-3 for module suggestion. Simply describe the feature you're looking for and Crowdbotics will suggest a module. You can try this by selecting the tab that says “Suggest with AI” on the righthand side of the Storyboard Tool.
3. Model Builder
In the Model Builder, access the GPT-3 functionality by selecting "Add Model" in the right hand Data Models tab. From there, select the experimental function and simply type in a description of the model you'd like to build. This function can come up with fields you maybe didn't think of (that are related to what you asked) and it assigns the appropriate field types for each field. This speeds up the process of building individual models, meaning your app is up and running sooner!
We’re taking our time identifying these use cases within our own product, and we’re optimistic that GPT-3 will unlock new avenues for functionality that enhance our users’ experience. Even if its disruptive power falls short of our expectations, it’s clear that GPT-3 marks a huge step forward in our ability to integrate AI into human interfaces.