We interviewed Varun Shenoy, a senior at Stanford who created a tool named GraphGPT. It transforms natural language input into a knowledge graph, using only JavaScript (React), HTML, and CSS.

devmio: Could you please introduce yourself to our readers?

Varun Shenoy: I'm Varun. I'm a Stanford senior studying a mix of electrical engineering and computer science.

devmio: You created GraphGPT, a tool that converts unstructured natural language into a knowledge graph. Can you tell us a little bit about what it does?

Varun Shenoy: GraphGPT is a system that uses GPT, a language model developed by OpenAI, to convert unstructured natural language text into a knowledge graph. A knowledge graph is a graphical representation of entities and their relationships.

devmio: How is GPT used in this case? How can it be used to make graphs?

Varun Shenoy: GPT is used as a natural language processing tool to understand the relationships between entities mentioned in the text and convert them into a structured format for representation as a knowledge graph.

devmio: Are there currently any limitations to GraphGPT, and how do you see them being amended in the future?

Varun Shenoy: There are currently limitations to the accuracy of GraphGPT's output, as language models like GPT are not perfect and still have room for improvement. These limitations could be amended in the future with further development and refinement of the language model. I'm also currently working on engineering a prompt that is faster and allows for assembling much larger graphs.

devmio: What are some of the most intriguing applications of this technology that you have come across? And how could it be advanced further?

Varun Shenoy: Graphs are everywhere and are an incredibly powerful data structure. I've talked with people applying GraphGPT to data indexing, code diagramming, and medical ontologies.

devmio: You only used JavaScript, CSS, and HTML for your project. Did you use any JavaScript frameworks or is it all vanilla?

Varun Shenoy: I'm using React.

devmio: On your website, you list a lot of projects. Can you tell us what you are currently working on?

Varun Shenoy: I'm a full-time student, so I’m spending most of my time heads down on schoolwork or hanging out with friends. In my free time, I'm exploring how large language models can be used to edit images, enable better search experiences, and help users better understand their own data.

devmio: The potential threat to jobs posed by the advancement of AI is often a topic of discussion. What is your viewpoint on this issue? Which professions are most at risk and what are the potential employment prospects?

Varun Shenoy: The current AI tools are not good enough to replace people, and I think as they get better, they will help people become much more efficient and powerful workers. AI allows for workers to focus on the most important parts of their trade and not worry about the menial stuff.

devmio: What development in AI/ML would you like to see in the near future?

Varun Shenoy: I want more options. Currently, everyone is using OpenAI's released models because they're the best. I don't like the idea that one company can monopolize this market. Hopefully, more companies will release their own models for public use in the next year.

devmio: Thank you so much for taking the time to answer our questions.