National Cloud Database Day

How has AI impacted developers’ experiences with cloud databases?

The democratization of AI has reduced the barrier to entry for developers to use AI technology within their applications. Developers have access to LLM and large multimodal marketplaces in the cloud, and open source datasets to train their own models. A distributed scalable cloud database with vector search capabilities is integral to the development and successful deployment of GenAI applications.

Developers are looking to augment their existing applications with AI capabilities. A major challenge with AI models is hallucinations. An AI application powered by retrieval-augmented generation (RAG) can query a vector database for relevant contextual data and send the results to an LLM to improve the accuracy of responses. Developers expect a unified data platform that, in addition to supporting vector search capabilities, is capable of full-text search, regular queries, and data analytics that can be used in combination with vector search.

Can you share any interesting or innovative ways developers can use GenAI to enhance the user experience or functionality of apps?

Developers can easily augment the experience of their applications by incorporating features like AI-powered chatbots that enable users to interact with applications using natural language and to get accurate responses to questions in an intuitive natural language format. Developers can build smart and adaptive guidance to simplify complex user workflows and repetitive tasks on behalf of users. Enabling advanced search capabilities across various data types and combining them with regular search filters further customize the result set.

While we discuss GenAI applications in the context of the cloud, decentralization of AI through edge AI applications is critical for the scalability of AI applications and for alleviating data privacy concerns. Data generated and analyzed at the edge allows developers to build highly responsive and reactive applications. Developers can also build personalized experiences by leveraging data generated and stored locally on devices.

How has the emergence of GenAI technologies like LLMs and code generation tools influenced developers’ approach to software development?

In 2023, developer productivity got a boost with GenAI-powered tools and applications like ChatGPT, GitHub Copilot, etc. In a recent survey posted by GitHub, developers who used GitHub Copilot completed the assigned task 55% faster than the developers who didn’t. These tools alleviate mundane and repetitive tasks like writing boilerplate code for unit tests, learning a new language (e.g., “translate my Java method foobar into Kotlin”), or onboarding to a new platform or technology (e.g., “show me a C code snippet that reads and writes data into platform X” or “provide an implementation of this interface definition”). It can be used for generating a draft of API documentation, documenting code, etc. Rather than completely taking over the developer’s tasks, AI coding assistants are used as paired programmers so developers and AI assistants work in concert to generate higher quality code at a faster pace.

How do you think GenAI will impact the roles of developers in the future, and what steps should developers take to upskill or adapt?

I would caution against developers from becoming too reliant on these tools and letting them do all or most of their work. It can stifle innovation, creativity, and critical thinking. Oftentimes, the results from GenAI tools aren’t the most efficient, accurate, or even up to date. Sometimes, there is additional security and compliance risk, such as the results being copyrighted. AI will continue to play a critical role in developer productivity as long as developers understand the limitations of these tools and exercise good judgment of how to use them as efficient assistants.