As the field of generative AI (GenAI) continues to evolve, data remains both its lifeblood and its greatest challenge. At the TechCrunch Disrupt 2024, DataStax CEO Chet Kapoor and other industry leaders addressed the issue of data overload in GenAI, emphasizing the requirement for strategic, data-focused initiatives that prioritize quality and practical utility over sheer scale. With generative AI still in its early phases, businesses are learning data access, which can lead to inefficiency and inaccuracy.
The Problem with Scaling Too Quickly
Kapoor pointed out that, while unstructured data is essential for AI, companies that dive head first into GenAI with a lot of data often end up overwhelmed. This is especially true for organizations that try integrating AI across various departments without a clear plan.
"Start small," advised Vanessa Larco, a venture capital company NEA partner. She proposed working backward, identifying the specific problem, and then locating and utilizing only the data required. This strategy avoids building a data "mess," in which irrelevant information slows down the AI.
Focus on Real-Time, High-Quality Data
Another major takeaway from the conversation was the value of high-quality real-time data. George Fraser, CEO of data integration platform Fiveran, underlines that companies should target immediate issues with correct, up-to-date data instead of gathering vast, unwieldy data sets. This method will help conserve resources and allow the team to concentrate on solving present problems instead of planning for the theoretical future scalability.
Fraser observed that failed attempts frequently represent the actual cost of innovation. As a result, using tailored data to solve today's problems decreases the danger of costly mistakes and missed steps caused by scaling too quickly using GenAI.
GenAI’s Early “Angry Birds” Era
Kapoor compared this stage of GenAI development to the early days of the mobile revolution, comparing today's GenAI applications to "Angry Birds"- enjoyable and promising but not yet transformative. Kapoor claims that while current GenAI applications might be useful for minor tasks, they haven't significantly impacted daily lives. However, he believes that 2025 will be the "year of transformation," and businesses will go from experimenting with AI in little ways to using it to change their operational trajectory.
Industry leaders agree that generative AI has a lot of potential. However, businesses should start small and concentrate on specific issues, utilizing accurate, high-quality data to prevent data overload. This realistic approach enhances outcomes and prepares GenAI for more impactful applications in the near future.