Hi, my name is Su Yeon Lee. I am a freshman studying Computer Science in the School of Engineering and Economics at Wharton. Two weeks ago on February 16, I participated in Penn’s Generative AI (GenAI) Hackathon, a 24-hour event hosted by Contrary for students to learn, build, and create GenAI-based applications. Our Moso Deal Team led by Nae Nae Chairatchaneeboon (MD) has been sourcing and seeking out innovative startups employing software technologies—GenAI being one of them—so I wanted to share my experience and notable projects from the hackathon, as well as broader insight for emerging startups to leverage its capabilities across diverse industries.

Before diving into the hackathon, for those unfamiliar: 

What is Generative AI?

Generative AI refers to artificial intelligence models capable of generating new content, such as text, images, audio, or video, that was not explicitly programmed into them. These systems learn from vast amounts of data to create realistic and tailored content almost instantaneously, while also iterating and improving on the generated output. This has profound implications across various industries, as it can significantly accelerate processes that involve content creation, whether it's writing software code, crafting marketing materials, or generating high-resolution medical imagery.

From Tourism and Education to Business Productivity

One of the standout projects was "Madame," which won the best overall award in the hackathon. This application leverages GenAI to create personalized stories and visual depictions of landmarks based on the user’s geographical location and preferences. By seamlessly integrating text generation (Baseten’s Mistral & Stable Diffusion), speech synthesis (ElevenLabs’ Turbo v2), and image generation models, "Madame" offers an immersive and personalized experience for exploring new places.

The "Philadelphia Tutor Assist" project highlighted GenAI's potential in education by generating tailored visual and audio content to help educators effectively communicate subject matters to students. Many EdTech companies have already been working to integrate GenAI into their products for personalized learning experiences. For instance, during my internship at an SF-based EdTech firm, I contributed to developing an AI chatbot that engaged students in discussions about reading materials. Beyond assessing comprehension, the chatbot dynamically formulated questions to foster active interaction, making learning not just effective but also stimulating and relevant.

Another project that stood out was "Sitdown," which leverages GenAI to streamline collaboration and productivity in the workplace. By automatically summarizing work updates, generating descriptions of websites and content visited, and visualizing team-wide workflow, this tool aims to replace unnecessary meetings while facilitating more effective communication and coordination within organizations. 

What can Generative AI Offer?

One notable project was "Dear My Diary," which integrates a GenAI model with a diary interface to offer teenagers a supportive virtual companion for journaling and receiving guidance — from the perspectives of a peer, parent, or counselor (depending on how the model is prompted). While making spaces like counseling more accessible, this also raises concerns about deploying GenAI in sensitive domains. The lack of control over the AI's responses, potentially harmful, highlights the need for robust content regulation, clear usage policies, and guidelines. As investors evaluate companies in GenAI, they should seek evidence of responsible development practices and monitoring processes, mitigating risks and positioning themselves for long-term success.

The common thread across these vastly diverse projects is the power of Generative AI for personalization and optimization. Historically, many offerings were designed as a one-size-fits-all solution, providing a singular experience to a diverse customer base. However, by integrating GenAI capabilities, companies can tailor their offerings to individual preferences and needs, resulting in more engaging, relevant, and ultimately satisfying experiences for their customers. This effect was previously glimpsed in many EdTech companies, which tailored learning materials and curriculum to individual students, thereby maximizing the educational benefits derived from their existing content.

Another key value proposition of GenAI lies in its capacity to automate certain time-consuming activities, enhancing operational efficiency and productivity. As highlighted by McKinsey's analysis, current generative AI and other technologies can automate up to 60 to 70 percent of work activities that consume employees' time today. The  "Sitdown” from the hackathon exemplifies this, automating workplace progress updates. It bridges the gap between managers and employees while allowing employees to focus on their main tasks, driving productivity. Another project that grabbed my attention was “Scisketch,” a software that automatically generates illustrations in scientific papers for researchers. It was fascinating to imagine how it could enable researchers to channel their efforts towards advancing their scientific endeavors, rather than grappling with the intricacies of illustration software. Investors seeking opportunities in the GenAI space should pay close attention to where these values (personalization and automation) are not yet being fully captured, and identify companies that are positioned to unlock this potential. 

‍

Concluding the GenAI Hackathon…

While working on my own project (if curious, play our game for training future prompt engineers…: promptle.onrender.com), I had the opportunity to witness firsthand the incredible potential of Generative AI technology across various domains. From enhancing educational experiences to revolutionizing business productivity and unlocking new avenues for creative entertainment, the projects showcased at the event were a testament to the transformative power of this cutting-edge technology.