How Anyone Can Build Meaningful AI Without Code - Ep. 283

Noah Kravitz welcomes Shania Levin, co-founder and CEO of Impromptu AI, to discuss building meaningful AI without code. Shania shares her background in business and computer science, including her experience at Google and eBay, which led to the founding of Impromptu AI. The company focuses on improving AI output accuracy for non-technical users, inspired by technology developed with Dr. Sean Robinson that achieved 98% accuracy.

The conversation highlights the importance of trust in AI technologies and the initial challenges faced by the Impromptu AI team, including manually assisting clients in creating AI applications. They developed a specialized AI builder after identifying common infrastructure needs across clients, primarily targeting large and mid-sized enterprises looking to transform legacy systems into AI-native companies.

Shania discusses technological innovations like adaptive context engines and infinite memory, which facilitate AI integration into existing code. The company emphasizes client education and employs a co-build model to foster collaboration between experts and client teams. She notes the significance of context and memory in AI, especially as code becomes more commoditized, and introduces custom data models for tailored responses.

Addressing the gap for first-time generative AI builders, Shania explains that while basic chatbots can be quickly created, scaling for enterprise use involves governance and control. She advocates for AI accessibility, promoting its democratization and sharing examples of diverse applications, from financial literacy to sustainable product development.

Shania highlights how AI can enhance recycling efforts by integrating systems and providing operational recommendations. She notes the necessity of custom data for unique applications and discusses the evolution of her daily operations with AI. Impromptu employs a mixed code approach, utilizing both no code and pro code, and leverages NVIDIA's CUDA libraries for performance, enabling faster product iterations and instant feedback.

The conversation touches on model benchmarks and task success, with users defining their own success metrics. Impromptu's system offers Manual and Automatic Optimization modes, achieving high task accuracy with minimal manual intervention. Shania stresses the importance of building trust through "provable AI," ensuring transparency in decision-making processes and addressing privacy concerns.

Shania reflects on her experiences as a woman in tech, discussing barriers to entry and the rapid changes in generative AI. She emphasizes the need to empower young individuals to pursue their ideas without technical limitations and advocates for a broader understanding of computer science education, focusing on critical thinking and problem-solving.

The discussion also highlights the significance of big picture thinking in technical roles, especially in product management, and the distinct skill of understanding AI-generated code. Shania and her co-speaker identify substantial opportunities for AI to revolutionize industries, enhancing accuracy and addressing challenges like climate change with the right focus and resources.