Last week I attended ODTUG KScope24. I lost track of count, but my first attendance was way back in 2011 in Long Beach California, and ever since (with some small interruptions) I have visited almost annually. Anyway, this year’s conference was in Nashville Tennessee, and I once again had the honor of presenting a session.
This year I presented a session on UI by AI, which was focusing on AI (or better GenAI) solutions that help you create User Interfaces. It was technology agnostic, although the second part of my presentation touched on Oracle APEX’s AI Assistant and Oracle Autonomous’ Select AI solution. Audience was (as far as I could see) enthusiastic to see something new. Imagine, many people hadn’t heard of GIT, Figma, V0, not had they ever used ChatGPT. Also, Oracle’s Select AI feature was unknown to many in the audience.
After introducing some key concepts of AI and GenAI I and an explanation of the differences between the two to I went into the Impact of GenAI on the Software Development Lifecycle.
Clearly GenAI has big impact on SDLC. It can enhance Code Review, by AI-driven code analysis, automated bug detection and suggestions and Improve code quality. Also with Automated Testing it can be helpful by (for instance) the generation of test cases. With this integrated in Continuous integration and delivery it will ensure faster feedback loops. Imagine what GenAI can do regarding documentation; It can generate comprehensive documentation and thus allow for easier onboarding for new developers.
But the biggest impact is most probably with Automated Code Generation. There are many tools to help you with this and it increases development speed and for instance reduce repetitive coding tasks. Also Design Assistance is very valuable. Think of the automated generation of design prototypes, with consistency in design elements and faster design iterations!
The main part of my presentation was about how GenAI can help you developing UI’s. I started by showing several examples such as ChatGPT, V0 and Quest.ai . All of these have their pro’s and cons. Depending on your own knowledge and your developer base either of these tools (and many others) can add considerable value. Quest.ai for instance Generates React.js code from your Figma designs. It integrates with popular design and dev tools like Figma, GitHub, etc. It supports major UI libraries like Material UI, Chakra UI, etc. Very good here is that it generates clean, extensible, production-ready code, where you are in full control as you can download the code or push it straight to your GitHub.
Next, I showed a demo of GitHub Copilot which seamlessly integrates with Visual Studio Code, Visual Studio, Vim, Neovim, and the JetBrains suite of IDEs. It improves Coding Speed, Boosts Developer Productivity, enhances Code Quality and Reduces Coding Errors. Even with the simplest JavaScript / HTML code, Github copilot was able to come up with suggestions for improvement and explanation of code. Always very nice to show this to the audience.
The final part of the presentation was dedicated to Oracle Technology, to be more specific, Oracle APEX and Oracle Select AI.
With the release of version 24.1, Oracle APEX has now the option to work with the APEX Assistant. This Assistant can help you with writing code and explaining code. It can give you (based on natural language prompt) SQL and PL/SQL suggestions, but also help you with JavaScript. The other thing it can do is create an application for you, based on natural language. So during a demo I asked it “I would like an application showing all employees and departments, and please add a salary chart to the application” It seamlessly came up with this application. It will work for you too (as long as you have EMP and DEPT in your schema of course).
Finally, I showed a cool demo of Oracle Select AI. That is, after explaining Select AI and how to set it up, I showed how it can be used in an APEX application. Details on select AI can be found in a previous post on this blog.
Key takeaways here should be that even when GenAI can be very helpful and content producing tasks will become largely automated you will always be responsible! If you do not have any domain of coding knowledge of whatever it is that GenAI creates for you, it can be wrong on so many levels. Also, never forget about the ethical considerations. AI can have bias in data and algorithms, and with GenAI keep authenticity of generated content and intellectual property issues in mind.
All in all, a nice session where I got positive feedback during and after the session. If you need any help with UI by AI please feel free to reach out. I am always willing to help you out.
